He runs a solo plaintiff personal injury practice, bringing in about $750,000 a year in fees. That’s his portion, after client recoveries and settlement disbursements are handled. He returns every inquiry within ten minutes and has built his firm’s reputation on that promise. He’s proud of this standard, and it’s well deserved. On my best days, I also return calls within ten minutes and keep track of those days.

The firm in this example is a composite based on engagement work, and all numbers are rounded.

Scope of the Agreement

The case count and response speed are both real. According to Clio’s 2024 secret shopper study of 500 firms, only 40% answered the phone, and that number rose to 52% with callbacks. Returning every inquiry within ten minutes puts this firm far ahead of the average.

The problem comes when people see the 140 as a direct measure of revenue. Each signing led to a written agreement, which is important. ABA Model Rule 1.5(c) says a contingent fee agreement must be in writing and must explain how the fee will be calculated, so he knows the formula for every case at signing. The actual fee is decided as the case moves forward, owner hours add up, and the final payment might come one to three years later.

So, the case count shows how many commitments intake made. Using it as a measure of financial return combines two different timelines into a single number.

Identical Timing, Varied Financial Outcomes

Let’s look at two cases he signed in the same week with the same type of agreement. The numbers here are just examples to show how it works.

The first case is a rear-end collision with clear liability, settled in nine months. The retained fee is $12,000, unreimbursed direct costs are $500, and he spends 30 owner hours on it.

The second case involves disputed liability and a surgical injury. It lasts 26 months, with depositions and two mediation attempts. The retained fee is $60,000, unreimbursed direct costs are $6,000, and it takes 260 owner hours.

The measure that applies to both:

Contribution per owner hour = (retained fee minus unreimbursed direct costs) ÷ owner hours on the matter

The first case brings in about $383 per owner hour. The second case earns five times the fee but comes out to about $208 per hour. So, the bigger fee paid him just over half as much for each hour of work.

When you do this for your own cases, make sure to keep the money categories clear. The settlement goes to the client, reimbursed costs are the firm getting back its advance, and the retained fee is what you use in the ratio.

Sources of Revenue

All 140 commitments were made before he could see the economics of the work, which is just how this practice works. Hyman, Black and Silver noted that screening in plaintiff personal injury practice starts with the first contact and continues as the case goes on. The first call opens the file, and the case reveals itself over the following months.

Fee income also comes in chunks. The same study found that a small number of cases can make up a large part of a practice’s fee income. Each year’s signings bring a mix of results, and just a few cases can make up a big part of the year’s income.

Timing is important, too. The $750,000 he collected this year was for work signed in past years, and this year’s 140 cases will pay out over the next two or three years. These two numbers come from different sets of cases. If you divide one by the other, you get a number that only makes sense on paper.

If you bill by the hour or use fixed fees, you’re making the same kind of commitment, just with the timing less obvious. Every owner puts in hours before knowing what the return will be, and contingency work puts that relationship right into the agreement.

Unanswered Questions in Dashboard Metrics

His dashboard tracks signed, open, closed, and collected cases, and those numbers are accurate. Practice management software can track time on contingency cases. My question for him, and for you, is whether your records link the retained fee for each closed case to the owner hours spent, all in one place, so you can calculate the ratio for every case as it closes.

The key point here is the uncertainty he faces from the day a case is signed until it’s finished. Each firm decides how to handle that uncertainty based on its own numbers.

Revisiting the 140 Cases

Signing 140 cases is a real achievement, and his intake team earned it. The front office does a great job attracting a certain type of case, and that’s what should be measured, one closed case at a time, in owner hours.

If you want a quick idea of where your firm stands, the Growth Intelligence Scorecard uses rough numbers you can recall from memory. It shows the gap between the cases you see now and the revenue that comes in later. The results are just a starting point and will be updated later, but they give a snapshot of your firm’s current revenue structure.

growthprolegal.com/scorecard

If you already track fees against owner hours in your contingency work, reply and let me know what you’ve found.

This letter is not legal advice or accounting advice.

The analysis in this letter is produced under Revenue Intelligence & Decision Architecture (RIDA), the proprietary economic discipline B.L. Sheets & Co. runs on and runs for the firms it serves. The doctrine, the case record, and the engagement formats are at blsheets.co.

Sources. Hyman, Black and Silver, The Economics of Plaintiff Side Personal Injury Practice, University of Illinois Law Review (2015); American Bar Association, Model Rule of Professional Conduct 1.5, Fees; Clio, client intake guide reporting the results of the 2024 secret shopper study; Clio, legal time and expense tracking. The firm is a composite drawn from engagement work. Figures are rounded and identifying details are withheld under the standing rule. The two case figures are illustrative.

He’s a solo criminal defense attorney in the Midwest, earning approximately $450,000 annually. He charges a flat fee for each stage of a case. He contacted me in March to discuss renewing his scenario modeling engine I had built for him, and the 90 open cases were mentioned early in the conversation as evidence of ongoing demand. This example is a composite, reflecting the economics of real practices with all identifying details removed. The 90 open cases are attributed to him.

The count is accurate; it’s likely he could identify each of the 90 cases. This figure consolidates both newly opened matters and those with multiple court dates into a single metric, with each matter counted once. The 90 open cases represent both the volume of incoming work and the duration each matter remains active. This number does not indicate the amount of time he still needs to dedicate to each file.

He interprets the 90 open cases as a measure of demand, which is partially accurate. The key metric for a sole proprietorship is contribution per owner hour, defined as the revenue generated from a matter divided by the number of owner hours invested minus fees and expenses. Revisiting a file without generating additional fees decreases this metric, while the open case count remains unchanged.

Same count, different weeks

Consider two months in his practice, both composite examples, each displaying approximately 90 open files at the beginning of the month.

In the first month, the majority of the 90 cases were opened within the previous 8 weeks, with few continuances. Upcoming court work primarily consists of initial appearances and conferences, representing first-time efforts on these files.

In the second month, the composition of the 90 open cases differs. Using the same starting date for each file, a greater proportion have been open since spring, with more cases continued one or more times. Upcoming court dates now include hearings for which he has already prepared, requiring repeated preparation for previously scheduled dates.

The composite time data for these two months is illustrative. In the first month, 11 of the 90 files were on a second or subsequent court setting, resulting in approximately seven hours of work on previously prepared files. In the second month, 34 files were on a second or later setting, with about 29 hours logged for repeat preparation. Stage fee collections remained around $37,000 each month, as his terms don’t permit additional stage fees for repeated work on the same stage.

Despite the same open case count and similar revenue in both months, approximately 22 additional hours in the second month were devoted to files previously prepared in the first month. Assuming other factors remain constant, this results in a lower contribution per owner hour, even though the open case count remains unchanged.

Why the old files hold the week

The National Center for State Courts analyzed approximately 1.2 million felony and misdemeanor cases from 136 courts across 21 states (data collected prior to the pandemic). Their findings indicate that time to disposition is influenced by two factors per case: continuances and hearings. In their model, each additional continuance extends the case by approximately three weeks. The report also highlights the cost to attorneys when preparatory work for a hearing must be repeated due to rescheduling.

These three weeks represent calendar time measured by the court. For the attorney, this results in the file returning to his workload.

Under a fixed stage fee arrangement, repeating work increases owner hours without generating additional revenue. Whether the original fee accounted for these extra hours depends on the engagement’s pricing structure. Some flat-fee agreements include such contingencies, but if the terms cover only one round of preparation and the court requires two, the additional effort goes uncompensated.

Delays may result from various factors, including actions by the prosecutor’s office, court scheduling, late discovery, client needs, or strategic defense considerations. Many continuances are appropriate and serve the interests of the client and due process. The primary concern is the cost to the owner, a decision made jointly with the client. Much of the case aging is beyond the attorney’s control, so the open case count may increase without corresponding new demand.

A significant challenge arises when older files require attention during the same week as new matters. When capacity is strained, the focus often shifts to new intake, which appears to be the primary source of workload. Files on their third or subsequent settings also consume substantial time, yet their impact is less visible in standard metrics.

What his reports can’t produce

His case management system and the court’s records contain extensive data, including court dates, file status, whether hearings occurred or were continued, and the reasons for continuances. Court data standards provide separate fields for open, inactive, and closed files, and record both the reason for continuance and hearing outcomes. This provides a comprehensive history of each file and how often it returns.

The time spent revisiting files isn’t captured in standard metrics, which is the focus of contribution per owner hour. Industry time metrics are typically based on billable hours. According to Clio’s 2026 Legal Trends Report (released October 26), projected utilization is 42% in 2026 (up from 39%), with the average lawyer recording 3.4 billable hours per day. In a stage-fee billing model, there is no billing entry for repeated work on the same stage, so these hours are excluded from such metrics. The key question for this firm is whether the continued presence of a matter on the calendar reflects the actual hours invested.

Monitoring case age relative to the applicable time standard, measured consistently from a single starting point, is essential. The court’s timeline begins when the complaint is filed, which may precede the attorney’s involvement, so choose and stick to a consistent start date.

If you want a first go at your own firm’s numbers then the GrowthProLegal scorecard is at growthprolegal.com/scorecard.

Look out for the new Growth, Scenario & Predictive Modeling engines at growthprolegal.com in November 2026!

This is general information only and should not be regarded as legal or accounting advice. The firm referred to is a composite and no client, court or case is in any way identified.

The analysis in this letter is produced under Revenue Intelligence & Decision Architecture (RIDA) which is the proprietary economic discipline that B.L. Sheets & Co. runs on and runs for the firms it serves. The doctrine, the case record and the engagement formats are all at blsheets.co.

Sources

• Clio, 2026 Legal Trends Report release (October 26, 2026)

• NCSC, Timely Justice in Criminal Cases: What the Data Tells Us (2020)

• NCSC, Model Criminal Continuance Policy and Implementation Guide (May 2024)

• NODS, Criminal Cases Recommended Data Elements (February 2025)

A solo immigration attorney earning about $500K annually reports that nearly 80% of her new matters come through introductions, and she has never needed to advertise. She presents this as evidence of her firm’s health. Two primary relationships drive most of these referrals: former clients who refer relatives for family petitions, and an HR director at a regional employer who introduces her to other employers needing assistance with employment visas.

She’s not alone in relying on referrals. Clio’s 2025 Legal Trends research for solo and small firms found that 59% identify referrals as their primary source of leads, compared to 27% of larger firms. This figure reflects the number of firms that trust referrals, but doesn’t indicate how much of any single firm’s work comes from this channel.

The number she trusts

The 80% figure is accurate and reflects new matters. It effectively answers whether people refer work to her. She also uses it to assess which types of work each source provides. A simple matter count doesn’t always capture that detail.

Several additional metrics support this figure. These include her share of collected fees and contribution by source, the amount remaining after delivery and acquisition costs. Most importantly, the share of her own hours each source requires is critical, as she is the sole attorney and her time is limited. We used 80% as a proxy for all these measures.

Same channel, different matters

Both referral streams come through introductions and often use flat fees. According to AILA’s 2022 marketplace study, 87% of its members bill primarily on fixed fees, making this a standard practice. The matters themselves require different approaches.

A family petition begins with Form I-130 and proceeds to adjustment of status or consular processing, depending on the relative’s location. This process is lengthy and involves ongoing communication with families. Employer matters use Form I-129, and H-1B petitions require a certified labor condition application before proceeding. These cases involve HR departments, each with its own schedule and definition of urgency.

Scope, fee size, the extent of paralegal involvement, and how often files require her direct attention all vary between these two types of matters.

It remains unclear which source best compensates her time, as her current reports don’t provide this information. Contribution per owner hour by source would reveal what each stream collects, minus delivery and acquisition costs, divided by her personal hours. When differences are measured, they reflect how the firm manages these types of work, not the referral sources themselves.

How the mix keeps itself

The same relationships continue to send similar matters. The mix of work remains consistent over time, without any deliberate decision to maintain it.

The cost of an introduction is also worth considering. While it appears as zero on the marketing line, and the 80% figure reflects this, an introduction that costs nothing to acquire can still require significant owner involvement. These hours count as delivery, which the referral percentage doesn’t capture. I once assumed a free lead was always beneficial, but experience has shown otherwise.

Counting matters identifies her busiest referral source. But the source that best compensates her time may differ, and matter counts will always highlight the source sending the most files.

What her reports can’t produce

Her records may link referral sources to billed or collected amounts, but they don’t track delivery costs or her personal hours. Contribution per owner hour requires both. Her current reporting doesn’t support a side-by-side comparison of the two streams.

Monitor the mix of work by source and watch how it changes over time.

For an initial assessment of your firm’s numbers, the GrowthProLegal scorecard is available at growthprolegal.com/scorecard.

This letter is general information and isn’t legal advice or accounting advice. The firm described is a composite, and no client or employer is identified.

The analysis in this letter is produced under Revenue Intelligence & Decision Architecture (RIDA), the proprietary economic discipline B.L. Sheets & Co. runs on and runs for the firms it serves. The doctrine, the case record, and the engagement formats are at blsheets.co.

Sources

• Clio, Highlights from the 2025 Legal Trends for Solo and Small Law Firms

• AILA, Hourly vs. Flat Fee Billing

• USCIS, I-130

• USCIS, I-129

• USCIS, H-1B Specialty Occupations

He c…

Read more

He couldn’t understand how he got here. He has a full calendar, and all of his invoicing is done.

The one move

The only number he looked at was revenue. At the old firm he generated $1.2 million a year, and when he left the book came with him.

He didn’t realize the significance of the firm his book was attached to before he left it. You know, the firm that paid him on time whether his clients did or not.

His book had been calculating two things at once. Not just the work itself, it also had a balance sheet that was built to cover any gap between finishing the matter and the client actually paying the invoice. The old firm handled the second one for him. He only saw the one number. The revenue he generated.

If you read my work often, you know what comes next. The number that matters in a solo or small firm is contribution per owner hour. Measured when the cash arrives, after the cost of waiting for it, and minus expenses.

Same book, two balance sheets

His clients are companies billed hourly on net terms, so the gap between finishing the work and receiving the money is built into the book.

Clio’s 2025 benchmarks put the median realization lockup at 43 days, the time between doing the work and billing it, and the median collection lockup at 32 days, the time between billing and payment. The median total lockup is 93 days. Each median comes from its own distribution, so the total is the 93 Clio reports directly. On a $1.2 million book, 93 days of lockup holds about $306,000 of finished work outside the bank at any given time. Each ten days of lockup holds about $33,000.

At the old firm that $306,000 sat on someone else’s balance sheet. The firm’s cash, its credit line, and its draw schedule absorbed the delay, and it stayed out of his view. At his own firm, the same amount starts on day one, and his own balance sheet carries it.

What moves the timing

His payroll runs on a two week cycle and his collections run on each client’s schedule. The two clocks run independently.

A book weighted toward a small number of corporate clients concentrates timing along with revenue. One client’s slow quarter becomes one missed payroll, where a larger book would show a dip in the average.

The interest on the credit line and his own deferred draw are both real costs of holding the gap. Both land in the overhead account. They’re detached from the matters that created the gap, and the clients whose timing caused the delay look clean on paper.

What the reporting leaves out

In the first quarter, contribution per owner hour turns negative once you calculate carrying cost. Even with revenue identical to the old book. The clients are the same. The rates are the same and the hours are the same. The number changes only when the date the cash arrives is part of the calculation.

The P & L and the bank balance both show cash is tight. But neither can tell you which matters cause the delay.

The zone worth watching is cash coverage of fixed costs against lockup. I’m naming the zone and leaving the threshold for a later pass, once more quarters are in.

A single point projection gives false certainty, so the exposure belongs in a range. The $306,000 is a median spread evenly across a year still ahead of him. The actual figure depends on how slowly his clients move this month and by how much. Cash flow is decided in the tail of the distribution, and the median describes the middle.

Pull one ratio

Pull one quarter first. Fees, interest, and draw are in the books, and owner hours come from the calendar.

Contribution per owner hour after carrying cost = fees collected on the book this quarter, less the variable cost of delivering the work, less the interest on borrowing used to cover the wait and the owner draw deferred to cover it ÷ owner hours worked this quarter

Run it by client first, then by matter class. Use the same definitions next quarter, then run it again. The direction across quarters carries more information than any single quarter.

Where this goes

This issue identifies the mechanism and the zone. It leaves the reserve, the size of the credit line, and the collections policy open, since Stage 1 is incomplete and a partial remedy on an incomplete read adds risk.

Revenue gets earned when the work is done and spent when the cash arrives, and contribution per owner hour gets measured at the second moment.

If you want to see where your own firm stands before guessing, the Growth Intelligence Scorecard is free. It works from figures you can give in ranges from memory, so the read is provisional with reconciliation to follow.

growthprolegal.com/scorecard

This letter is not legal advice or accounting advice.

The analysis in this letter is produced under Revenue Intelligence & Decision Architecture (RIDA), the proprietary economic discipline B.L. Sheets & Co. runs on and runs for the firms it serves. The doctrine, the case record, and the engagement formats are at blsheets.co.

Sources. Clio, Law Firm KPIs and Benchmarks (2025), for realization, collection and total lockup medians. The attorney is a composite drawn from engagement work. Figures are rounded and identifying details are withheld under the standing rule. The exposure figures apply a median to a single book and show a mechanism. They leave the right reserve or credit line for a one-owner firm to that firm’s own numbers.

He has a firm specializing in estate planning and elder care in southeast Michigan, and the firm earned $715,000 last year. A few weeks before that he had implemented a five minute phone screening process. In February the firm conducted 24 consultations and took on 11 cases.

He regarded the new figure as evidence that intake had improved. The report he was looking confirmed his view and the figure was correct. What it implied, however, was still open to interpretation.

What the rate counts

The formula: Number of matters/Number of consultations

This ratio increased for two different reasons: either the quality of the consultations improved or the firm changed the kind of inquiries that count as one, and in both cases the dashboard showed the same line.

The execution aspect is genuine. According to Clio’s 2025 Legal Trends for Solo and Small Law Firms, conversion increased by 10% amongst firms that used electronic signatures and by 7% among those that used text messaging. These figures come from vendor data and suggest that streamlining the process can increase conversion numbers.

The screen is located on the other side of the fraction. Each inquiry it rejects passes over the denominator without having a consultation, and the numerator loses only those individuals who actually signed up. If inquiries that had signed up at a rate lower than the average are turned away, the percentage increases even though the number of people on the calendar decreases. If instead those who signed up at the average rate are turned away, the percentage remains the same as before. Sorry, that’s a lot of math.

The term “close rate” refers to more than one figure. According to the documentation on Clio’s dashboard, three of these are defined: the number of matters that have been signed compared to all inquiries, the number of matters that have been signed compared to the number of held consultations, and the number of matters that have been accepted compared to the number of matters created. Each of these figures begins to be counted at a different stage in the process, and one may increase while the others remain unchanged. The screen could be perfectly suited to his practice, since the percentages show the movement and leave the cause to him.

Walk two matters

Select two from the eleven.

The first one is an estate plan which was referred to and signed during the initial meeting, while the second is an elder care case that started with a web inquiry and required two consultations as well as a lengthy call with the family before it could be signed. The rate counts one signature for each of these cases and no further.

When buyers are hiring an attorney, they usually do their research and call several. According to Martindale-Avvo, 11% of legal customers chose the first lawyer they contacted and 78.9% of those who made a hire had contacted more than one, which shows what a buyer who is comparing does by arranging a second consultation. The fact that the signature is delayed tells something about the buyer and leaves the quality of the case to be judged in the rest of the record.

If you look at these two matters in turn, they end up being separate. One of them gathered more after accounting for delivery costs. The other required more owner hours to produce and complete. The conversion rate remains the same for both of them at 46%.

What moves the rate

Before the consultation is even touched upon, two forces have an effect on the number.

The first is the denominator: a rate will go up when the company changes who it counts as a consultation, even if the consultation itself remains just as good as before.

The second factor is source mix. According to Clio’s 2025 report, referrals are the main source of leads for 59% of sole practitioners and small firms, meaning that a firm which receives a higher proportion of referrals will have a higher blended rate even if the meeting it held last year remains the same. A firm that receives a broader range of cold inquiries, on the other hand, may see the rate drop yet still end up with more cases than before. Such changes could either benefit or disadvantage the firm, and the rate is shown as being the same in both cases.

The missing read

The third element is the owner’s own time, which is concealed within the percentages. A consultation uses up the same limited hours as a delivery does, and the owner loses that hour regardless of whether the prospect agrees to proceed or not. According to Clio’s 2025 benchmarks, the average lawyer’s utilization is 38%, which is equivalent to three billable hours in an eight hour day, the remaining five hours being spent on intake and administration and other tasks. The close rate assigns no value to those hours.

The data is often already in the system; Clio Grow is able to show conversion rates and revenue according to the source of the lead, and Clio Manage keeps track of time and the amounts collected for each matter. Each report is accurate within its particular area and all of them are awaiting being brought together. The combined figure is obtained by dividing, for each intake stream, the amount collected after variable costs by the total number of owner hours that the stream used, and the standard conversion view omits this.

Pull one ratio

Start by withdrawing a quarter; the fees and costs are listed in the practice management system and the owner’s hours are taken from the calendar.

Contribution per owner hour, by intake stream = fees collected on the stream’s retained matters this quarter, less the variable cost of delivering them ÷ owner hours spent on the stream’s consultations and on delivering those matters this quarter

First check it with the lead source and then with the matter class. Use the same definitions next quarter and then run it again.

More information is contained in the movement across the quarters than in any one quarter alone. If the contribution of a stream per owner hour is increasing, then it is paying for the calendar space that it occupies. If the stream’s contribution is staying the same or falling, it is drawing on the calendar at a rate which the close rate has been kept out of sight.

This brings us to the area that should be given a name: the intake qualification rule, which determines which sources and matter types get owner consultation time. Locate where your rule is set out and read it next to the ratio. The decisions that come after these are the firm’s own.

Where the signature sits

The figure relating to intake was accurate. For each hour that he spent on a particular stream the amount collected is shown in a number of columns to the right, and it is that number which determines the year.

The signature appears as the final line in the intake report and as the first line in the economic report.

If you want to get an initial view of your own figures, the Growth Intelligence Scorecard can read out your company’s revenue structure using ranges of figures that you remember. It takes about four minutes to complete in your web browser, and the result is provisional, with a reconciliation to follow, showing which revenue streams are worth looking at more closely.

growthprolegal.com/scorecard

The letter should not be regarded as legal or accounting advice.

The analysis contained in this letter has been carried out using Revenue Intelligence & Decision Architecture (RIDA), the proprietary economic methodology which B.L. Sheets & Co. both applies and for which it provides services to its clients. Further information regarding the doctrine, the case records and the engagement formats can be found at blsheets.co.

The sources include Clio’s 2025 Legal Trends for Solo and Small Law Firms; Clio Law Firm KPIs and Benchmarks (2025); the help documentation for the Clio Grow Dashboard, the Clio Grow Reports and the Clio Manage Productivity Reports; and Martindale-Avvo’s ‘Three Things to Know about Legal Consumers’. The firm presented here is based on previous engagement work. The figures have been rounded and specific details have been omitted in accordance with the standard rule. The Clio conversion figures are derived from data provided by the vendor. As for the external research, it confirms the mechanism in place and leaves it up to each one-owner firm to determine its own intake figure.

The last time we spoke, the calendar had made the decision for her. DUI and misdemeanor files kept coming in, with enough felony work mixed through the month to wreck whatever schedule she thought she had. The firm was collecting around $430,000. She was handling the legal work, the client calls, and the decisions that arrived five minutes before court.

The associate was billing on schedule. Her monthly production looked steady. Add up the year, and her billings were over double her salary. The invoices said the hire was working, and the owner was working more than she had before the hire.

She wasn’t calling to complain about the associate. She was trying to reconcile two records that seemed to describe different firms. The spreadsheet showed added capacity. Her calendar showed evenings spent reviewing motions, fixing drafts, and returning calls from clients who started with the associate and ended up asking for the lawyer whose name was on the door.

The numbers looked fine. She was tired enough to wonder which one of them was lying.

What the billings mean

The associate’s billings were real. She had opened files, moved cases, appeared in court, drafted motions, and recorded time. Nobody needed to explain those numbers away. The question was what the owner had assumed they meant.

A producing associate can add revenue and consume owner hours on the same file in the same week. The invoice carries the associate’s production. It has no place for the owner’s review, correction, instruction, or the client call that moves upstairs after fifteen minutes.

That omission changes the read.

A hire can produce enough billings to look self-supporting while drawing from the same owner calendar it was hired to release. Salary against billings tests payroll coverage. It doesn’t test whether the owner bought back any capacity. Those are different transactions and the standard report records only one of them.

Walk one file

Take a misdemeanor file from that week.

The associate handles the intake and drafts the first motion. The owner reads it before it goes out. Some drafts need a few comments. One comes back with the argument headed in the wrong direction, so the owner rewrites most of it after dinner. No timer is running. There’s no client bill waiting for that entry, and the practice management system doesn’t interrupt at 9:20 p.m. to ask whether the owner has quietly become the second associate on the file.

Then the client calls. The associate answers the first set of questions. The client asks for the owner when the conversation reaches the plea decision. The owner takes the call because the client hired her firm, the judgment is hers, and handing off work never handed off responsibility.

The professional rules recognize that structure. ABA Model Rule 5.1 requires a lawyer with direct supervisory authority to make reasonable efforts to ensure that another lawyer follows the rules. Its commentary specifically includes systems that ensure inexperienced lawyers receive proper supervision.americanbar+1

Supervision is part of producing the legal work. It still disappears from the production report.

By Friday, the associate has recorded eleven billable hours on the files we traced. The owner has spent four hours reviewing, correcting, instructing, and finishing the work those eleven hours started. The firm reports eleven new hours of production because the four owner hours went missing. They came out of the same calendar the hire was supposed to free.

The missing cost

Every report the firm runs answers some version of the same question: did the associate produce? Hours billed. Fees billed. Collections credited. Payroll paid.

None of those reports was built to answer what that production cost the owner directly. Supervision wasn’t set up as a matter anyone could bill. The time system doesn’t ask for it, so the P&L receives no owner-hour cost to place beside the associate’s output.

Management research treats attention as a scarce organizational resource, not a free input. Research published in Management Science found that a manager’s limited attention capacity affects where that manager can spend time under pressure. Onboarding research has also described new hires as an initial draw on productivity because they consume training time and coworker attention before reaching full production.

Neither study sets a law firm ratio. They support the mechanism sitting in this owner’s calendar. The associate’s output requires another input to produce it, and that input belongs to the owner.

This is how a hire can clear the visible tests and still leave the owner’s economics worse, or how the right hire can run at a supervision ratio the firm never intended. The salary gets placed against the associate’s billings. The comparison comes out favorable because the owner’s hours never reached the other side of the ledger.

The firm hired to buy back owner time. Eight months later, it still hadn’t measured how much owner time the hire used.

Pull one ratio

Pull one month first. The numbers are already in the practice management system, although the owner time may have to come from the calendar and memory.

Owner hours spent supervising per billable hour the associate produces = owner review, correction, client-transition, and meeting time on the associate’s files this month ÷ the associate’s billable hours this month

Use the same definitions next month. Then run it again.

There’s no universal answer for where this number should land. Criminal defense work, experience, matter mix, and the owner’s review obligations all change the ratio. A first-month number also means little on its own. Training is supposed to use time. The direction carries the information.

A ratio that falls across the months describes an associate producing more work for less owner supervision. A ratio that holds steady describes a continuing draw on the owner’s calendar, even while the associate’s billings rise. The guardrail belongs where associate production and owner supervision draw keep moving in opposite directions on the same hour.

That’s the zone to name. It doesn’t decide whether the associate stays, whether another person is hired, or whether the work should be supervised differently. Those decisions belong to the firm and its own numbers.

The ratio supplies the missing owner-hour input. Put it beside the contribution from the associate’s work and the firm can read what the hire adds per hour of owner supervision consumed. Until then, the billings describe production without describing the capacity used to create it.

Half the transaction

A hire’s own numbers can be correct and still describe half the transaction. The other half sits on the owner’s calendar. It doesn’t appear on an invoice because nobody built the invoice to carry it.

The associate may be producing. The payroll may be covered. The owner can still be spending four hours to create eleven that the report assigns to someone else.

If you’ve never measured what a hire costs you in your own hours, the Growth Intelligence Scorecard reads your firm’s revenue structure from the numbers you already carry. About four minutes in your browser, no meeting.

growthprolegal.com/scorecard

This letter is not legal advice or accounting advice.

The analysis in this letter is produced under Revenue Intelligence & Decision Architecture (RIDA), the proprietary economic discipline B.L. Sheets & Co. runs on and runs for the firms it serves. The doctrine, the case record, and the engagement formats are at blsheets.co.

Sources. ABA Model Rule 5.1 and its comments on supervisory responsibility; research on managerial attention published in Management Science; MIT Sloan Management Review research on the productivity cost and ramp period of new hires. The owner is a composite drawn from engagement work. Figures are rounded and identifying details are withheld under the standing rule. The outside research supports the mechanism. It does not establish a supervision threshold for a one-owner law firm.

She had a laminated fee sheet. I noticed it before anything else on the desk, and I’ve thought about it a few times since. You don’t usually laminate something you plan on changing anytime soon.

A family law owner, roughly $500,000 collected the year before, all of it fixed-fee. She specialized in uncontested divorces and custody modifications. She priced them as packages off that sheet, which she built herself and has only adjusted twice in five years. She dropped time entries three years ago. The firm doesn’t bill hourly, so what would she enter? Right?

So I asked her what a custody modification pays per hour once you count the file review, the filing fees she fronts and sometimes never sees again, the two continuances that ate an afternoon each, and the client calls at 9 pm.

She didn’t know.

And she’s sharp. She knows her collections. She can quote every package on that laminated sheet from memory, and she out-earns just about every lawyer she went to school with. She didn’t know because nothing in her system would tell her, and nobody ever built anything that would.

I’ve spent twenty years inside law firm economics, a good stretch of it sitting in the CFO, pricing, or finance seat at firms too small to justify a full-time one, and that conversation ends the same way every time. The owner decides pricing analysis is a BigLaw thing. Too expensive. Built for a scale she’ll never see. So, she filed it mentally next to the rest of what big firms do that has nothing to do with her.

I had an associate pull every US law firm posting for a dedicated pricing title they could reach this week. I asked that it be run twice, once across the trailing twelve months, once restricted to what was live in the last ninety days. Forty-seven distinct postings over the year. Thirty-two of them live in the last ninety days. The median posted midpoint is just over $142,500, and $161,250 if you look only at manager and director seats.

She’s right about the money. She wasn’t in any position to hire a full-time pricing analyst.

What they’re actually buying

The postings repeat so closely you could write one of them from the other 46. Alternative fee arrangement design, 25 of the 47. Pricing support for RFPs and pitches, 24. Building the financial model, 23. Matter budgeting and then watching the matter against that budget, 21. Profitability analysis by client or by matter, 17. 43 of the 47 describe joining a pricing function that already exists. 2 describe building one from scratch. 2 don’t say which.

So this is a function that’s been sitting inside large firms for years, getting staffed harder this year. Every one of those duties rests on the same thing underneath: cost captured at the matter level, hours and cash both, set against what came back.

Now the floor, which matters more than the salaries. Of the firms disclosing a headcount against a dedicated pricing title, the smallest is a 225 attorney shop where the pricing manager reports straight to the CFO. The rest run 500, 750, 800, 1,100 lawyers. The floor only drops when pricing stops being its own title and turns into one duty inside a finance seat. Even then, the smallest firm anywhere in the set is around 45 attorneys, hiring a director of finance to carry budgeting, partner profitability, pricing strategy, and investment decisions, all of it, for $120,000 to $240,000 base. (Read that duty list again. That’s four jobs.) The only posting in the entire research that touches small firms is a fractional CFO practice advertising to serve about thirty of them at once, from the outside, part-time, with pricing mentioned once, in a line it shares with margins and cash flow.

Nobody is hiring a pricing manager for a $500,000 family law practice. That was never going to happen, and the research confirms it. The research also shows that the question the pricing manager gets hired to answer keeps mattering all the way down. It just stops being anyone’s job somewhere along the way, which is a quieter problem and a worse one.

Why a fixed-fee firm assumes this doesn’t apply

Time tracking and hourly billing got welded together so long ago that dropping one feels like it should retire the other. It doesn’t, and that’s the misunderstanding. Margin on a matter runs on the full cost of producing the matter, and the fee that came in is a different number.

Here’s the layer I’m working on. Cost to produce a matter is the hours the file took, valued at what the hour is worth, plus what went out in cash alongside it. Filing fees. Service of process. An expert or a transcript when the matter needed one. The direct admin time spent scheduling and preparing the file, separate from the attorney’s own hours. Full overhead allocation, spreading rent and software and insurance across matters, is a deeper accounting exercise and not this letter’s job. I’ve been the guy asking about filing fees for twenty years now. It has never improved a dinner party once. Recover your fees!

A fixed fee tells you what came in, and on its own it tells you nothing at all about what went out.

Which is why this sits in front of the pricing question. A firm can have well designed fixed fees, a clean sheet, a pricing philosophy, and still have no way to answer what any one package cost to deliver. The people getting hired this year exist to close that gap at a size where closing it is worth a salary. Everywhere below that, same gap, nobody assigned.

No formula this week.

So one question, instead. Can you take the matter type that fills most of your calendar and tell me what it cost the firm to produce this year? Cost means the hours the work took at what those hours are worth, plus the filing fees and other direct costs that went out alongside them.

Yes means the number those firms are paying $161,250 to compute is inside your reach right now, without hiring anybody.

No means you have a decision. Whether to build that capture with your own numbers. And a newsletter has no business making it for you. What the no tells you is that the firm can’t currently ask a question the market just spent a year proving is worth six figures to answer. Worth knowing before any pricing decision gets stacked on top of it.

BigLaw worked out what guessing costs and decided the guess ran more expensive than the salary. That’s the entire discovery.

The guess is still on the table at every size below the one hiring for this. What’s gone is the excuse. The question applied before anyone could afford to answer it, and it still applies after a firm decides it can’t.

Don’t know what your busiest matter type costs to produce? The Growth Intelligence Scorecard reads your firm’s revenue structure from the numbers you already carry. About four minutes, in your browser, no meeting. growthprolegal.com/scorecard

This letter is not legal advice. It is not accounting advice.

The analysis in this letter is produced under Revenue Intelligence & Decision Architecture (RIDA), the proprietary economic discipline B.L. Sheets & Co. runs on and runs for the firms it serves. The doctrine, the case record, and the engagement formats are at blsheets.co.

Sources. Two internal snapshots of active US law firm job postings, collected 2026-09-14: dedicated pricing titles over a trailing twelve-month window (47 distinct postings) and the same search restricted to postings live in the trailing ninety days (32 distinct postings), plus a broader finance and operations title search carrying pricing responsibility over the same ninety days (14 distinct postings). Figures cited are the employer’s own posted ranges except where noted. The owner in the opening is a composite drawn from engagement work, rounded and anonymized per the standing rule.

She was right about the calendar. Every hour was spoken for. The conclusion she came to, that time is the constraint on the firm, is the claim we’re about to test. We call this letter The Billable Hour, so it seems right to run the breakdown here.

Two owners, one full week

We’re going to take a look at two owners with identical weeks.

The first is the owner on the phone. Trace her hours file by file, and they land at about 45% of collections running through a single carrier. This is the work that ended the agency lunches mentioned earlier. The stopped lunches are what let the referral work go quiet. The quiet referrals increased the dependence on the single carrier. Her hours are full because one claims department’s assignment queue decides what her week looks like. The calendar is where concentration shows up.

Let’s compare that to a fixed-fee estate planning practice. Same gross over the same time period. Trace his hours, and they pool in package work priced in 2022, and nobody has ever calculated what those hours pay per owner hour. The collection is where the number moves. At solo and small firms, about 14% of billable work never reaches an invoice, and roughly a tenth of what is invoiced goes unpaid (Clio Legal Trends Report). His hours are full because a growing share of the week earns an effective rate nobody has looked at.

Same felt problem, but the constraint that binds each firm is different, and so is the fix. Get that wrong, and you execute the right answer to the wrong problem, at full cost and zero effect. I’ve watched owners do it. I’ve done it on my own firm (not recently), which is the kind of thing you’re not supposed to admit in a letter about diagnosis.

Where the hour comes from

Here’s the mechanism. The owner’s hour is the surface where everything that binds a one-owner firm shows up, because there’s one calendar and everything in the firm clears through it. A demand problem shows up as hours spent chasing work. A pricing problem shows up as hours that collect less than they billed. A concentration problem shows up as a calendar one client controls. The hour registers all of it and identifies none of the causes.

Which is why tracing what consumed the hour, matter by matter, is the diagnostic act, and treating the hour itself as the problem is a product category. The time-management shelf sells calendars to lawyers, but a better calendar just reorganizes the symptom.

The two numbers to pull this afternoon

Two numbers decide the read, and your reporting produces neither. Both come from figures you already have.

Contribution per owner hour, by matter type = (collected on that matter type - the direct cost of producing it) / owner hours it consumed

Practice management software reports billed hours and gross revenue. This number lives in the gap between those two reports, which is how the work that fills most of a week can pay the least per hour for years without anyone noticing. After collection, the ranking of what a week is worth rarely matches the ranking of what fills it.

Concentration = the largest relationship’s trailing twelve months of collections / the firm’s trailing twelve months of collections

You can name the anchor from memory, the client or the carrier. Stating the share is different work, and the share is the number a lender, a bad quarter, or one change of adjuster will eventually state for you. There’s a share of collections above which a relationship stops being a client and starts being the firm’s structure. That zone is real. Where the line sits for your firm depends on your own decomposition.

Pull the anchor’s collections for the last twelve months and divide. Then pull one matter type, the one that filled the most hours, and run the first formula. Do those two and you’ll know more about the firm than the full calendar has told you all year.

The top 3% of solos gross $600K to $1 million and above (Clio Legal Trends), and nearly all of them arrive there on full calendars. The hour is evidence. A full week tells you a constraint exists, just not which one.

If the concentration number came back higher than you expected. The Growth Intelligence Scorecard reads your firm’s revenue structure from the numbers you already carry. About four minutes in your browser, no meeting. growthprolegal.com/scorecard

This letter is not legal advice and is not accounting advice.

The analysis in this letter is produced under Revenue Intelligence & Decision Architecture (RIDA), the proprietary economic discipline B.L. Sheets & Co. runs on and runs for the firms it serves. The doctrine, the case record, and the engagement formats are at blsheets.co.

Sources. Clio Legal Trends Report: realization and collection figures (14% of billable work unbilled at solo and small firms, roughly a tenth of invoiced work unpaid) and the solo revenue distribution (top 3% grossing $600K to $1 million and above). Both owners are composites drawn from engagement work, rounded and anonymized per the standing rule.

I was guilty of doing this myself when I started out. I justified it by giving it a name: client relations. I thought it was a cool name. But it’s actually just a rate cut. Being the numbers guy, I was embarrassed to say I never ran an analysis on what it cost me.

The billed numbers and the hours are real. The mistake is in the meaning of the total. It sits there as a record that a price was attached to some of the hours. It doesn’t say anything about the hours that were cut before the price on the bill was set or about the portion of the price the client decides later to leave unpaid.

Clio’s Legal Trends Report puts numbers on both of those stages. At solo and small firms, about 14% of all billable work never gets invoiced. Then, what does end up on the bill, another 10% on average gets left unpaid. Stack that, and $1k of work becomes $860 on a bill, which only collects and deposits $775.

She wishes she charged what was on her rate card. But in reality, what she was actually charged was “banked per hour.” Both of those numbers were set by her, but only one of them was set on purpose.

Here’s what the billing report can’t tell her. Two matter types, same $27k, and the report says they’re equal. Composite firm, rounded numbers.

Matter A is her flagship. Hourly litigation is the work she built her firm on.

She worked 62 hours at $500. $31k. After her “review,” she billed 54 hours. $27k. Then, after the client does their own edit (haircut), she ends up with $24.3k banked. Per owner hour = $24.3 / 62 = $392

Now Matter B is the work she tries to avoid at all costs. Fixed fee: 10 matters at $2,700 each.

She worked 40 hours and billed $27k. That’s just the fee. Nothing for her to edit. She collects the full $27k at signing, and her banked per owner hour is $27k / 40 = $675.

***Before anyone reads that as a case for fixed fees, run the same ten matters in a month where just 2 of them go sideways. Worked 70 hours. Billed and collected $27k banked per owner hour is $27k / 70 = $386

It’s the same $27k a third time, but this time the fixed fee is paying less than the litigation. Here’s why I am an advocate for hourly billing. The report didn’t move, right? The hours did, and the report doesn’t have a column for hours against a fixed price. I have been writing the case against fixed fees for years. Even in the”value pricing” craze we’re in now. Here’s my argument: a fixed fee just relocates the write down. On hourly work, the write down usually comes at bill review, where at least you can see yourself doing it. On a fixed fee, it happens the day you quote $2700 for 10 matters that were never going to cost the same, and every hour past the estimate is a write down you take in advance without a prebill to catch it. Every matter is different. Every single one of them. The fixed fee prices them as if they are all the same, all of the time.

Leaving the pulpit now.

The 40 hours won the first comparison here, and the form doesn’t get the credit. The reports treat them as equal. What she would see, if she runs it, is the one she likes the least and the one she protects the most trade places on nothing but the hours. The price is left hanging because she never tied the hours to it.

Where the $4,000 went

The write down always happens in 3 places.

First is at her desk while the time goes in. She takes a call and enters it as .8 because they had a little small talk during the actual 60 minutes of the call. Never counted, so never counted as a loss. The small talk was part of the call. Bill it under “client relations” if you want, but bill it.

Second is bill review. The Friday example above. The owner is selling herself short to avoid client conflict she fears, and in my experience, it wouldn’t actually happen. In reality, it’s a discount offered after the work was done and never solicited by the client, so it doesn’t carry any of the information a discount would normally carry. Please don’t discount. The missing information in this case is what the owner was afraid the client wouldn’t pay for. Again, no column for that on the report.

Third is at collection: the invoice on a closed matter comes in short, but she’s okay with that because the matter’s over and the relationship isn’t, and she doesn’t want to reopen a closed file to argue about $900. But that’s not okay. It becomes a habit and routine. And without changing your rate card, your prices go down.

Add the three up, and the firm has quietly repriced itself 22% below the rate card. It still says $500, and she keeps believing $500, and she prices the next hire and the next matter off a number that stopped being true.

There’s a second force I wrote about in an article last week: delay. Delay compounds this one. The longer the work sits, the more of it dies in old WIP.

The number to pull this afternoon

Here’s what your reporting should show if the billing system was set up by someone who knew what they were doing and cared about your success: Billed by month. AR aging. A realization percentage for the whole firm. What it can’t show is banked per owner hour, by matter type. The inputs are all there. Your calendar has the hours, billing has the bill, the bank has the cash, but nothing currently connects any of it.

Worked hours = hours recorded before bill review, by matter type.

Banked = cash received on those matters, after the last short payment.

Banked per owner hour = banked / worked hours

Run it for one month and one matter type. Then run it on the matter type you like the least. I’ve shown dozens of owners how to set this up. Once they ran it on one matter for one month, none of them stopped. All of them implemented this practice wide.

There’s a point when bill review stops being “client relations” and becomes the firm repricing itself. Where that point sits for a given firm and which matter type crosses it are stage 1 RIDA™ findings. This letter doesn’t make it.

We are currently building the engine to do it for you, though. Coming in Q4 2026.

Remember this: billed is what your firm asks for. Banked is what you get back. In a firm with one owner and a fixed number of hours, that gap is a pay cut the owner takes at their own desk, sometimes in tenths of an hour, on the work that fills your calendar, at a rate nobody printed. The matter type you’re protecting is often the one that pays you the least once those two numbers meet, and the lack of reporting that keeps them apart is why it stays that way.

The hour you billed was paid at a rate you never agreed to.

If the number you pulled is smaller than you thought. The Growth Intelligence Scorecard reads your firm’s revenue structure from the numbers you already carry. About four minutes, in your browser, no meeting. growthprolegal.com/scorecard

This is not legal advice. This is not accounting advice.

The matter figures are composites drawn from engagement work, rounded and labeled as such. Market figures are from the Clio Legal Trends Report, paraphrased and attributed.

Sources. Clio Legal Trends Report (realization and collection benchmarks).

The analysis in this letter is produced under Revenue Intelligence & Decision Architecture (RIDA), the proprietary economic discipline B.L. Sheets & Co. runs on and runs for the firms it serves. The doctrine, the case record, and the engagement formats are at blsheets.co.

These days I can type a few usable words a minute on a good day, with errors I can’t reliably fix afterward. So I talk.

Every essay, post, and comment I’ve published in the last several years started as my voice in a room. Dictation software turns the sound into text and gets most of it right, which is doing well given how I’m prone to mumbling. Claude and Grammarly now handle the repairs after that. Punctuation mostly never arrives in a dictated transcript, and when my speech gets bad the software starts guessing at words, so somebody has to put the sentences back the way I said them. The arguments and the numbers and the stories are mine, and so are the jokes that don’t land.

I’m explaining my setup because of what happened next.

On July 30, LinkedIn shipped a button that lets any member flag a post as AI generated. Flagged content gets reduced distribution. The account holder gets no notice, no stated standard, no appeal, and no undo. I know about the undo part with more certainty than I’d like, because the man who flagged one of my comments later apologized in public, said he’d take me at my word, and went to remove his flag. He came back a few minutes later to report that the platform gave him no way to do it. He was decent about the whole thing, which somehow made it worse to watch.

Since early August my impressions are down about 50-70% from July. My weekly article reached only 24 people in the first 90 minutes last week beyond the subscribers who get it by email. I’ve published daily for ten years and I sit near 13,000 followers, and nobody at the platform has been able to tell me which lever made those numbers stop adding up. There’s a case open through LinkedIn’s Disability Answer Desk, and the people there have been professional with me so far, though completely silent for two weeks now.

While the review runs I went back to the only thing I control, my own process. I’ve spent thirty years telling owners to quit arguing with outcomes and fix the system underneath, so I took my own advice. Here’s what I write under now.

I dictate long and messy. I speak the piece the way I’d argue it across a table, restarts and tangents included, and I don’t stop to fix anything mid-flow. Back when I corrected as I went, everything came out in short uniform sentences, which is exactly the pattern that gets called machine writing now. So the mess stays in until the end.

I save the raw take. The audio file and the untouched transcript get dated and archived before any tool touches them. If you dictate your writing, do this. If anyone ever presses a button on you, you’ll be holding a recording of yourself saying the words.

The correction pass repairs and doesn’t rewrite. My instructions to the tools are narrow. Fix transcription errors, fix punctuation, fix sentence boundaries, and where the audio was unclear, mark the spot instead of guessing. I had to learn to tell the tools no rephrasing, in those words, because they want to help, and what their help does over a few passes is make my paragraph sound like everyone else’s.

Grammarly gets spelling and agreement and nothing else. The clarity and engagement suggestions stay off. Those features exist to move every writer toward one register, and I’ve watched what happens to that register on LinkedIn lately.

I read the draft back out loud before it publishes. Anywhere it sounds smoother than I talk, I roughen it. The test is the one I’ve used on client memos since the nineties. Would I say this sentence across a table to a managing partner I’ve known ten years, or does it sound like a keynote.

And every piece carries at least one real number. A figure from an actual matter or an actual month, rounded and anonymized, but real, because that’s the one thing generic content can’t produce, whoever or whatever wrote it.

Two things I decided against. I won’t run drafts through AI detectors, because the research puts their accuracy near a coin flip, and chasing their approval would bend my voice further than any correction tool has. And I won’t seed fake typos. I have the audio, and performing sloppiness would be its own kind of fake.

And because this letter runs on the economics of a one-owner firm, I’ll run the read on myself. The protocol isn’t free. Archiving the takes, running the tighter correction passes, and reading every draft back out loud adds time, call it an hour a piece, and I publish multiple essays and articles weekly across several properties. The case adds more. Screenshots, correspondence, analytics pulls, and the hours I’ve spent proving I wrote my own sentences. August has taken something like a working week out of my calendar so far. I run one-owner businesses, so those hours came from where they always come from in a firm like yours, out of the work that pays. Contribution per owner hour is the number this letter exists to protect, and mine went down this month for reasons that had nothing to do with my clients or my craft. When an enforcement error lands on a one-owner firm there’s no associate to hand it to, so the owner carries it, usually in the evening, after the paying work.

The part that stays unfortunate is that none of this should be necessary. A disabled writer shouldn’t need an evidence chain to prove he wrote his own sentences, and a platform shouldn’t attach permanent distribution consequences to a button with no standard behind it. I asked one of the flag movement’s advocates which level of AI involvement earns a flag, since he’d published a scoring system for it that morning. He told me, “it’s in the eye of the beholder...” I’ve thought about his answer every day since, because he wasn’t wrong about how the button works. That is how it works.

Twenty years ago the tremors took the putter and the pen. They don’t get the writing.

-Sheets

The calendar was full. That part’s a fact, and nothing in this letter argues with it. The error is in what full gets taken to mean. A full calendar is a volume reading. It says the hours were consumed, and it says nothing about what consumed them or what any of them paid, so the hiring reflex ends up reading a volume number as a value number. It answers the constraint the owner can feel before anyone’s named the one that governs. The hour is where every constraint in a solo firm presents. Where it lives is a separate question, and the hire usually gets made before anyone asks it.

Two hires, same rough budget, opposite directions. The reflex version is the associate at $110K, and she produces billable hours, which is the point. She also consumes owner hours, which the spreadsheet never books: training, file review, supervision, the redo on the early drafts, and that draw runs heaviest in exactly the months when the owner has the least to give, which is the condition that triggered the hire in the first place. The other version is an operations and intake manager at $58K. Zero billable hours. Per the Clio Legal Trends Report, the average lawyer captures about 3 billable hours in an 8-hour day, and the other five go to admin, intake, and running the firm, so this hire is aimed at the five. Get two of them back a day, across 240 working days, and that’s roughly 480 owner hours a year, and at an effective $250 to $350 per collected owner hour it’s somewhere between $120K and $170K of capacity at the owner’s own rate, off a $58K salary, before she’s answered a single call.

The associate math deserves the same treatment, because the version that approves the hire prices hours at the rate card. Target 1,400 hours at $300 and the projection reads $420K against a $110K salary, and the hire approves itself. Billed is an intention. At solo and small firms about 14% of billable work never reaches an invoice, and roughly a tenth of what’s invoiced goes unpaid, per the same Clio data. Associate work in its first year realizes below the firm’s average on top of that. Run those against the projection and the $420K lands somewhere in the high $200Ks to low $300Ks collected, and the owner’s supervision hours come out of that return at the owner’s own rate. The associate can still be the right hire. The decision needs to run on the collected numbers, and the rate card projection isn’t one of them.

None of this shows up in standard reporting, which is why the reflex survives. The time system tracks the billable hours and ignores the five that aren’t. The practice management report shows gross by matter, the bank shows collected in aggregate, and nothing in between computes what each kind of work pays per owner hour after collection and direct costs. Supervision time isn’t booked anywhere at all. So the decision defaults to the one number that’s always visible, which is the full calendar, and the calendar can’t answer it.

The structure, stated once. Owner hours earning below the firm’s average contribution per owner hour are the candidate zone for delegation, and the hire that clears that zone can outperform the hire that adds production. Where the boundary sits is the firm’s own number, from its own decomposition, and this letter doesn’t know it.

Sometimes the most logical next hire is the one that makes current revenue sustainable, and that hire bills nothing.

The Growth Intelligence Scorecard runs the read in general form: four minutes, in your browser, from numbers you already carry, and it names the constraint your own figures point to. It’s free at growthprolegal.com/scorecard.

This letter is not legal advice, and it is not accounting advice.

The analysis in this letter is produced under Revenue Intelligence & Decision Architecture (RIDA), the proprietary economic discipline B.L. Sheets & Co. runs on and runs for the firms it serves. The doctrine, the case record, and the engagement formats are at blsheets.co.

The firm was in its third growth year, roughly $900,000 billed, and the quarterly marketing review was impressive. Impressions up. Cost per lead down 18%. Consultations at a record. Gross billings at a record. Four numbers on one page, each one accurate, each one trending the right way. The agency’s rep walked the deck in twenty minutes. The retainer continues for another quarter.

Then the owner reconciled the operating account. The quarter closed within dollars of the same quarter last year. This doesn’t add up.

Nobody in that review had an explanation. They couldn’t point to anything in that report that could provide one.

The Turn

The number is real. What it implies is wrong.

Every metric in that report focuses on how much was spent. Impressions and clicks price what the ad auction charged for attention. Leads and consultations measure how much demand was created, with every inquiry counted the same regardless of its value. Gross billings record what the firm intended to collect. Three meters, all bolted to the same side of the transaction: the side where money and hours go out.

The report supposedly shows ROI. What they’re actually presenting is “spend” packaged as “return.” This is the entire problem. And improvement in the metrics does nothing to fix it. A cheaper lead is a better price on the spend. It doesn’t carry any information about what came back.

Same Line, Different Money

Run two matter types off the same P&L. Both put $10,000 a matter on the gross revenue line, so the practice management report scores them identical.

The first is fixed-fee transactional work. It collects at 97% inside 30 days, and a matter consumes about seven owner hours once the drafting, the calls, and the closing are counted. The second is hourly work billed in arrears. It collects at 84% after 90 or more days, and a matter has eaten twenty owner hours by the time the file closes.

Same line on the revenue report. After collection and direct costs, the first pays more than $1,200 for each owner hour it ate up. The second pays under $400.

A gap that wide, running through a single revenue line, is completely missing from any of the agency's reports. It compounds, too. The second kind of work generates the most inquiries, so the record consultation quarter fills the calendar with work that pays a third as much per hour. The marketing engine gets better and better at buying the wrong matters.

Name your own pair. Every firm has one: the work that fills the week and the work that pays for it. They’re rarely the same work.

The Ladder

The mechanism is a ladder, and each rung throws away information the next rung needs.

Click. Lead. Billed. Collected. Contribution per owner hour.

A click doesn’t tell us if the visitor ever became a matter. A lead doesn’t tell us what the matter was worth, because volume counts a $2,000 inquiry and a $40,000 inquiry as one each. Billed doesn’t tell us if the money arrived, and that part is measurable: per the Clio Legal Trends Report, the average lawyer collects $910 per $1,000 of billable work performed, about 14% of billable work at solo and small firms never gets invoiced, and roughly a tenth of what does get invoiced goes unpaid. Collected doesn’t tell us what the return cost in owner’s hours to produce them.

The marketing report stops at rung two. The firm’s economics start at rung four.

The full map, one row per metric.

The join

Here’s why the metric gap survives year after year unmeasured. Contribution per owner hour requires a partnership among three systems that don’t communicate with each other. The practice management platform reports gross by matter. The bank records collected in aggregate. The hours, where they’re tracked at all, sit in a third system, and at fixed-fee firms they’re usually not tracked at all, which is why the effective rate goes invisible. No standard report performs the join. The one number that prices the owner’s most valuable and scarce resource is the one number no system produces by default.

The close

A vanity metric measures what was spent. A structural measure prices what came back. Growth that survives measurement is the only kind worth buying.

The candidate zone is the work paying below the firm’s own average per owner hour. The threshold is the firm’s number, produced by the firm’s own decomposition, and this essay doesn’t know it.

The Growth Intelligence Scorecard runs this read in general form, from numbers you already carry. Free, four minutes, in your browser: growthprolegal.com/scorecard

This letter is not legal advice and not accounting advice. Scenarios are anonymized archetypes with rounded figures.

Subscribe now

The analysis in this letter is produced under Revenue Intelligence & Decision Architecture (RIDA), the proprietary economic discipline that Legal Growth Intelligence runs on and runs for the firms it serves. The doctrine, the case record, and the engagement formats are at blsheets.co and growthprolegal.com

Four essays in this series read the seller’s numbers to the seller’s clients. This one reads them to the seller, because the owner of that agency is this newsletter’s reader in a different chair: a founder-led professional services firm, a fixed labor base sold by the hour, and an owner who is also the marketing department, since 79% of agencies have no one dedicated to their own marketing and 70% have no full-time salesperson.

The number the shop steers by is the pipeline dashboard. Leads created, proposals out, pipeline value against target. It’s the number the agency sells, and it’s the number the agency manages itself by, which feels like consistency.

Here is what the dashboard actually measures. Two things, in the same cell. The first is demand manufactured: leads generated, for clients and for the shop alike. The second is what that demand converts to in contribution once the labor base delivers it. The dashboard reports the first. The industry’s own books report the second, and they’ve been answering for four years. Billable utilization averaged 66.4% in 2025, down from 68.9%, the first time the industry has fallen below its own 70% minimum floor, with EBITDA margins dropping in direct correlation. Idle capacity is the demand panel telling the truth. The top stated reason agencies miss utilization targets is lack of client work. Demand binds the industry that sells demand.

Now the part the client essays couldn’t see: what the obligation costs. The average agency puts 7% of revenue into its own sales and marketing, and with no dedicated marketer in four shops out of five, the spend is invoiced largely in founder hours, the same constrained resource this newsletter prices for lawyers every week. The books show where it lands. Project margins average 35%; net margins arrive at 13%. Twenty-two points, consumed by overhead, inconsistent selling, scoping, and low-margin service lines, and the proof-of-product marketing sits inside that gap paying rent. At the top of the market, the average competitive pitch costs $204,461 in staff time and free ideas, and two in three of them lose. A founder-led shop runs the same event at its own scale: the proposal built across three unpaid weeks, the spec concepts, the discovery calls, all of it priced in the founder’s own hours, at the same one-in-three odds. Fee inertia deepens it from the other side: 64% of agencies plan to raise fees this year, and 16% actually raise them annually on existing clients.

The tactics run as credentials don’t close the gap. They widen it, because they consume the owner’s hours and the firm’s margin to manufacture a panel the buyer has stopped trusting anyway.

The proof of the alternative is in the same dataset. Agencies that narrowed their service mix grew 13% on average and posted 30% net margins. Agencies that expanded their offerings averaged 10% net. Same market, same year, opposite allocation. The winners didn’t outspend the category on marketing. They cut what they sold, which is a constraint decision, and the market paid them for it at roughly three times the industry’s average margin.

And the buyer side is enforcing the test whether or not the shop runs it. 60% of senior marketing leaders report spending less on agencies this year because of AI, and among teams that successfully adopted AI agents, 73% cut agency content spend, against 0% of teams that didn’t. After an average 8% headcount cut across agencies in 2025, Forrester forecasts 15% of agency jobs eliminated in 2026; WPP alone dropped from 108,044 people to 98,655 in a year. Forrester’s own analyst adds the delivery-side squeeze: 75% of agencies are absorbing the cost of AI work, and 6% have managed to monetize it. The honest counterweight belongs here, because the numbers support compression and nothing stronger. The agency count grew to over 71,000 in North America from 50,000 two years ago, ad spend rose 8.6% last year, and average revenue growth recovered to 7.5%. The category isn’t dying. It’s commoditizing, and a commoditizing demand lever punishes exactly one thing: an undifferentiated labor base marketing itself harder.

The measurement the dashboard can’t produce is the one this series handed the lawyer in Part 2, and it transfers without modification. What does each service line pay per owner hour after delivery cost and collection. What does each client pay, on the same denominator. The utilization report can’t answer it either; utilization is the agency’s version of the full calendar, a volume gauge wearing down an owner who has never seen the contribution gauge. Run the read and the answer decides the narrowing question with arithmetic instead of nerve: the service lines that survive are the ones that clear the shop’s own book, and the ones that don’t were never proof of product. They were load, priced as credential.

The read is the same one this property runs for its readers every week, and the arithmetic doesn’t care what the firm sells. The Growth Intelligence Scorecard computes contribution per owner hour from a firm’s own figures, in ranges, from memory, in about four minutes, and names the binding constraint. It was built on law firm economics; a founder-led agency is the same species with a different rate card. Free, at growthprolegal.com.

The series closes where it opened. The proposal and the survey describe the same companies, and now both sides of the table can read both panels. The seller’s clients learned to ask what a marketing dollar lands on before renewing it. The seller’s own books have been asking the same question for four years. It’s one discipline.

The Billable Hour returns to its regular run next Wednesday.

This issue is an economic diagnosis of a firm as a business. It is not legal advice and not the practice of law. It works from published industry data stated in ranges, is not a reconciliation, and is not accounting advice or a substitute for the firm’s accountant. It is not a valuation and not investment advice.

Subscribe now

Sources. SPI Research, 2025 Professional Services Maturity Benchmark, 403 firms (utilization at 66.4%, down from 68.9%; the 70% floor; four-year decline; EBITDA correlation). Parakeeto / Summit CPA (lack of client work as the top cause of missed utilization targets). Promethean Research, 2026 State of Digital Services, 119 agency leaders surveyed February 2026 (13% net against 35% project margins; $4.43M average revenue; narrowed mix at 13% growth and 30% net; expanded mix at 10% net; 7% of revenue to own sales and marketing; agency counts). SparkToro / Founder Focus, State of Digital Agencies 2025, 376 owners (79% no dedicated marketer; 70% no salesperson; fee inertia at 64% against 16%). ANA / 4A’s / Advertiser Perceptions, Cost of the Pitch, 2023 ($204,461 average; incumbent retention two in three). Typeface Signal Report, October 2025, 200+ marketing leaders VP and above (60% spending less due to AI; 73% against 0% among AI adopters). Forrester, Predictions 2026: Marketing Agencies (8% 2025 headcount cut; 15% 2026 forecast; reported holding company headcounts; 75% absorbing AI costs against 6% monetizing, per Jay Pattisall via The Drum). Guardrail figures: Forrester data via Ritner Digital (ad spend up 8.6% in 2025).

The number doing the work in that signature is the menu itself. The implied claim under every line item is that these services are what produces clients, and the proof on offer is that a company in the business of growth chose to sell them. The signature is being applied to the menu’s claim about causation.

Here is what the menu actually measures. Two things, in the same cell. The first is what the agency’s labor base can produce: the deliverables its headcount was hired to ship. The second is what produces clients. The menu reports the first and is silent on the second. There is one document in the file that reports the second, and it is the one the firm owner never sees: the seller’s own marketing budget.

The industry surveyed itself on that document, and the comparison can be run line by line.

SEO is sold as the core retainer, with standard guidance steering 45 percent of a law firm’s digital budget toward it as the highest-return channel, per the LEXGRO aggregation. In SparkToro’s survey of 376 agency owners, it is absent from the top drivers of the agencies’ own new business.

Paid search is sold as the fast lever, at legal click costs running $80 to $200 and an average of 13.4 leads to convert one client, per National Law Review and Martindale-Nolo figures. It is essentially absent from the sellers’ own acquisition stack, and the paid formats agencies do buy for themselves, networking events and industry awards, sit at the bottom of their own effectiveness rankings.

Social media management is sold as a standard line item. In the same survey, 20 percent of agencies say social media is no part of their own marketing at all, and the platforms dropped down their rankings as a source of business. The exception proves the shape: among agencies that do use social, 73 percent name LinkedIn most effective, and what works there is the founder writing under their own name, which is a different product from managed multi-platform posting.

Outbound lead generation is sold as intake campaigns and cold outreach programs. 59 percent of agencies have run it on themselves. 9 percent call it very effective. 33 percent call it not effective at all.

Content is the partial exception, and the honest version of this essay says so. Agencies genuinely use it. What they use is founder authority content and conference speaking that feeds the referral engine, produced by the founders themselves, since 79 percent of agencies have no one dedicated to their own marketing and 70 percent have no full-time salesperson. Volume content built for rankings, the version on the proposal, is a different product wearing the same name.

And the document has a bottom line. Asked what actually produces their new business, the sellers answer: referrals from past clients first by a wide margin, referrals from partner firms at 15 percent, founder content, and event speaking, which climbed from sixth to fourth in a year. In the broadest survey of the field, 93 percent of firms in the business of selling growth call their own growth engine weak. The engine that runs the category is relationships and reputation. The engine on the proposal is the six line items.

Read the seller’s allocation as what it is: the honest survey. An agency answering a questionnaire is describing itself; an agency spending its own money is revealing what it believes produces clients, at its own price, with its own margin at stake. What they buy for themselves is the belief. What they sell is the capacity.

The counterweight, because the honest claim is narrower than the satisfying one. The gap is incidence, not hypocrisy. Agencies buy against their own binding constraint, and for a referral-run professional service that constraint is relationships and reputation, so their budget goes there. A firm whose binding constraint is genuinely retail demand, high-volume consumer work in a market of strangers, may rationally buy tactics the seller does not buy for itself. The gap does not convict any line item. It convicts the menu’s claim to universality: the assumption, priced into the signature, that what the catalog contains is what any firm’s growth requires. That assumption is exactly what a binding-constraint read exists to test, and the sellers have never run one on themselves.

The reader can’t see the agency’s books, so the mirror can’t be checked directly. One question substitutes for it, asked before the renewal: which of these line items do you buy for yourselves, at your own price, and what did each one return. A shop that made its own allocation on evidence can answer in numbers, and the numbers will be interesting whatever they show. A shop that can’t answer is selling capacity in a growth costume, which is the same finding Part 2 reached from the other side of the table.

The two documents now have names. The proposal is the labor panel read out loud: what the headcount can ship, priced by the month. The seller’s own budget is the demand panel kept private: what the seller believes produces clients, priced with its own money. The series opened on the gap between those panels, and this is the gap completed. Before the next renewal, the number that prices any line item against the firm’s own economics is the one this series has been circling: contribution per owner hour. The Growth Intelligence Scorecard computes it in about four minutes, from your own figures, in ranges, from memory, and names the binding constraint. Free, at growthprolegal.com.

Next week, Part 5, the last in the series, written to the other side of the table: the agency owner, whose books show what the obligation to perform marketing costs.

This issue is an economic diagnosis of a firm as a business. It is not legal advice and not the practice of law. It works from published industry data stated in ranges, is not a reconciliation, and is not accounting advice or a substitute for the firm’s accountant. It is not a valuation and not investment advice.

Sources. SparkToro / Founder Focus, State of Digital Agencies 2025, 376 agency owners and consultants surveyed September to October 2025 (new-business drivers; partner referrals at 15 percent; speaking’s climb; social at 20 percent not part of own marketing; LinkedIn at 73 percent among users; outbound at 59 tried, 9 very effective, 33 not effective; 79 percent no dedicated marketer; 70 percent no full-time salesperson). RSW/US, 2025 Survey Report, Rolling Toward 2026 (93 percent of marketing services and professional services firms calling their growth engine weak). LEXGRO 2026 aggregation (the 45 percent SEO allocation guidance, cited as the sold-as figure, not endorsed). National Law Review 2025, Consultwebs 2025, Martindale-Nolo 2024 (legal click costs $80 to $200; 13.4 leads per client).

They earned different money. The transaction paid roughly $1,360 per owner hour. The dispute paid about $490. Same line on the P&L, nearly three to one apart in what the hour produced. Gross collections measure two things at once: what the work paid, and what the work consumed to get paid. The blend produces a firm-wide average that no actual matter matches, and the average is the number the owner trusts.

The number that governs is contribution per owner hour: what each kind of work leaves behind, after its direct costs, divided by the owner hours it consumed. In a true solo practice the collapse to the owner’s hour is definitional. Every hour the firm spends is the owner’s hour. There is no associate to absorb the 61-hour matter. The denominator is the owner’s life.

The calendar, meanwhile, is allocated by demand. Referral sources refer the work they have seen you do. Renewal work renews on its own schedule. A matter type that arrives steadily will fill the calendar whether it pays $490 an hour or $1,360, because arriving steadily is the only qualification the calendar checks. The transaction came once, through a broker. The dispute reproduces. Five years of this and the book has a shape that demand chose and the owner inherited. The firm’s highest-yield work can sit at the edge of its own book, admired and unexpanded, while the lowest-yield work compounds.

The spread also understates itself, because contribution is computed on collected dollars, and collected dollars run downhill from the rate card. Clio’s Legal Trends research puts the slope in figures. The average lawyer collects about $910 for every $1,000 of billable work performed. At solo and small firms, roughly 14 percent of billable work never reaches an invoice, and about a tenth of what is invoiced goes unpaid. The same research finds the average lawyer captures about three billable hours in an eight-hour day, and the median firm waits 93 days between performing the work and depositing the money. These leaks land unevenly across the book. The dispute with the difficult client bleeds the write-downs the transaction never sees. Measured per billable hour, the two matters sit three to one apart. Measured per owner hour, after the leaks, the gap widens, and the report that would show the widening does not exist.

That is the structural fact underneath the metric. Billing software reports hours by matter, where hours are tracked at all, and fixed fees tend to untrack them. The P&L reports collected dollars by client. Neither report divides what a kind of work collected by the owner hours it consumed, so the one number that ranks the book by what it pays for the owner’s time never prints. The owner knows the gross to the dollar and cannot state which third of the calendar funds the firm and which third the firm quietly subsidizes. Both facts live in numbers the owner already has. Fee, hours, direct costs, collections. The missing piece is a division the reporting was never built to perform.

A full calendar at $940,000 can hold a firm earning $1,360 an hour for a third of its time and $490 for the rest, and the gross will bless the whole thing. The gross is real. The question it cannot answer is which work earned it.

A disclosure. I built the thing this essay argues for. The new GrowthProLegal website and growth engines launched this week at growthprolegal.com after months of development. The Growth Intelligence Scorecard is live there now, free, and it stays free. It is a four-minute conversation, answered from memory in ranges, that adapts to the structure of your firm and names the binding constraint on its growth from four candidates: Acquisition, Conversion, Monetization, Capacity. The same answers produce the same read. It gives no advice. It names the constraint and stops. No card, no contract, no obligation.

Two deeper engines are in testing and will be released in August 2026. The Revenue System Diagnostic builds the ladder this essay describes, your kinds of work ranked by what each pays per owner hour, labeled provisional and unreconciled until an engagement reconciles it. The Growth Intelligence Blueprint drafts the provisional structural map that follows from the read. Full RIDA engagements are also now open at solo and small firm price points, with published rates and fixed scopes, on the site.

The essay stands without any of it. The division is yours to run on a legal pad. I built the machinery for owners who want it run systematically, and the front door is four minutes long.

The analysis in this letter is produced under Revenue Intelligence & Decision Architecture (RIDA), the proprietary economic model created by B.L. Sheets. The doctrine, the case record, and the engagement formats are at growthprolegal.com

This letter is general economic analysis. It is not legal advice, not accounting advice, not a valuation, and not investment advice.

The economics of a one-owner firm.

The number doing the work in that signature is the ranking. Position one, held, verified monthly. The logic runs: the firm is more visible than any competitor, visibility produces clients, the retainer protects the visibility. What could a better report even look like?

Here is what the ranking actually measures. Two things, in the same cell. The first is where the page sits when the results load. The second is what sitting there delivers: the click, the visit, the consultation request. For twenty years the two moved together closely enough that one number could stand for both. The report still prints the first. The second detached.

The size of the move comes from the field’s own instrument maker. Ahrefs studied 300,000 keywords and updated the numbers this February. When an AI Overview sits above position one, the top result gets 58 percent fewer clicks. Stretch the view across two full years and the picture gets simpler: on queries with an AI Overview, position one now delivers about a quarter of the clicks it did in December 2023. On queries without one, about half. The page didn’t move. The click did.

The mechanism is visible on any results page. The answer now renders above the rankings. A searcher reads it and leaves. Clickstream data covering the first four months of 2026 puts 68 percent of US searches ending without a single click to the open web, up from roughly 60 percent in 2024. When an AI Overview is present, the rate runs near 83 percent. In Google’s AI Mode it reaches 93 percent, and there the organic results are replaced entirely, so a ranking has no surface to appear on. Pew’s behavioral study of nearly 69,000 real queries found users clicked a traditional result on 8 percent of visits when an AI summary was present, against 15 percent without one, and clicked a link inside the summary on 1 percent.

The exposure isn’t evenly distributed, and it skews toward legal content. Informational queries go zero-click at 74 percent; transactional queries at 31. Legal search marketing was built on informational content, the what-happens-if and do-I-need-a articles that earned rankings for a decade. The content class that legal SEO leans on hardest is the class the answer box absorbs first.

Which produces the strangest artifact in the file: the report improves as the flow declines. Impressions rose roughly 49 percent since the AI rollout while clicks fell by nearly a third, because every rendering of an answer box counts the page beneath it as seen. The instrument isn’t broken. It’s faithfully measuring a panel that no longer routes the clients.

The field’s own numbers confirm it from the inside. Ahrefs, the company whose tooling produces a large share of the industry’s ranking reports, reports its own blog in a two-year monthly decline in organic clicks, with direct traffic overtaking organic as its primary source. The people who sell the instrument have watched it stop measuring their own flow, and the field’s most prominent voices now publish under the banner of zero-click marketing. The transition isn’t a prediction. The sellers already made it.

Now the counterweight, because the honest version of this essay is narrower than the dramatic one. US organic search traffic overall was down only about 2.5 percent year over year as of January. And the click-rate decline found a floor: Seer Interactive’s tracking shows the collapse bottoming in December 2025 and partially recovering into early 2026, leveling off at a new baseline. Search didn’t die. The click repriced, and the new price appears to be settling. The claim is instrument mismatch: a shop still selling rankings and traffic in 2026 is selling a gauge calibrated to a click path that no longer exists for a large share of queries. The queries exist. The gauge reads the wrong panel.

The measurement the reader’s report can’t produce is the decomposition. Of the impressions the report counts, how many rendered under an answer box. Of the clicks that remain, how many arrived from queries the AI layer has not absorbed. Whether the firm’s name appears inside the answers themselves, which is where visibility now lives for the absorbed queries. A shop that has made the transition can produce those numbers, because AI citation, on-page visibility, and brand search are measurable today. A shop that can’t produce them is reporting the field name and hoping the meaning holds.

The zone where a guardrail belongs can be named without setting it. A retainer renewed against a metric that no longer decomposes into client flow is spend justified by the instrument, not by the firm’s economics. Where that leaves any specific retainer depends on the firm’s query mix, its intake data, and what the remaining clicks are worth, and none of that is answered here.

The report is accurate. Every number on it is true. The renewal is being signed against what the numbers used to mean.

Next week, Part 4: the services on the proposal, against the services the agency buys for itself.

This issue is an economic diagnosis of a firm as a business. It is not legal advice and not the practice of law. It works from published industry data stated in ranges, is not a reconciliation, and is not accounting advice or a substitute for the firm’s accountant. It is not a valuation and not investment advice.

B.L. Sheets writes on the economics of founder-led firms, including revenue architecture, capital and transaction readiness, and pricing, margin, and retention. For law firm owners, the free Growth Intelligence Scorecard at growthprolegal.com reads the structure under your revenue in four minutes.

Sources. Ahrefs, February 2026 update of the AI Overviews CTR study, 300,000 keywords (58 percent reduction; two-year decline ratios), authored by Ryan Law and Xibeijia Guan. SparkToro / Similarweb clickstream analysis, June 2026 (68 percent US zero-click, first four months of 2026). Similarweb zero-click tracking, 2024 to 2025. Pew Research Center, July 2025, 68,879 queries (8 percent against 15; 1 percent inside summaries). Bain-Dynata Generative AI Consumer Survey (83 percent zero-click with AI Overviews). Semrush (AI Mode 93 percent; intent split, 74 percent informational against 31 transactional). BrightEdge (impressions up roughly 49 percent since the AI Overviews launch while clicks fell nearly 30 percent). Graphite via Search Engine Land, January 2026 (US organic traffic down 2.5 percent year over year). Seer Interactive tracking, updated April 2026 (December 2025 floor, early 2026 leveling). Ryan Law via The Marketing Meetup, February 2026 (Ahrefs’ own blog decline; direct overtaking organic).

The number doing the work in that signature is the calendar. Some owners use its cousin, the monthly billings total. Either way the logic runs the same: the firm is busier than last year, the spend preceded the busyness, so the spend is working. The retainer gets renewed against fullness.

Here is what the calendar actually measures. Two things, at the same time, in the same cell. The first is volume: how much work arrived, got opened, got scheduled. The second is what each hour of the owner’s time earned after collection. The calendar reports the first and is silent on the second. A week can be full and the firm can still be earning less per owner hour than it did when the weeks had gaps.

The number that governs the renewal is the second one, and it has a name: contribution per owner hour. The fee collected, minus the variable cost of producing the work, divided by the owner hours the work consumed. Not revenue. Not the fee on the engagement letter. What the work paid for the hours it took, after write-downs and collection.

Marketing is a demand lever. It moves the volume number, and it can move it hard. What it cannot do is change what the firm earns per hour once the work arrives. On a one-owner firm the owner’s billable hour is the binding constraint. There is no associate to absorb overflow, so every new matter competes with every existing matter for the same constrained hours. When the hours are full, an incremental matter displaces something, and what it displaces is chosen by the intake funnel, not by contribution.

This is why Monetization binds before Acquisition on a firm like this one. Demand pushed into a structure that underprices the owner’s hour adds unbillable time, write-downs, and load, and the busier the calendar gets, the faster the underpricing compounds, because the matters that arrive fastest tend to be the ones priced to arrive fast.

The industry’s own numbers describe the result. Aggregated 2026 survey data puts 74 percent of law firm marketing budgets in low-ROI activities, a figure that replicates across at least three sources asking the question differently. Clio’s attribution work found 25 to 35 percent of legal marketing spend wasted to poor tracking alone, before any question of whether the leads were worth having. The utilization benchmarks run the other side of the ledger: 65 to 80 percent of hours billed is the sustainable band, and above roughly 85 percent, quality and durability decline. A full calendar on a solo is frequently a firm operating above that band, paying a retainer to stay there.

The owners closest to the constraint moved first. Solos cut marketing budgets at the highest rate of any firm size in the most recent spend survey, 24 percent. Read as revealed preference instead of retreat, that is the segment with the least slack concluding fastest that the spend was buying something other than what the invoice said.

Put the two panels together and the structural claim falls out. Where Monetization binds, more demand makes the firm busier and poorer at once. Each marketing dollar that lands another underpriced matter converts owner hours into load. The corrective direction, in that state, is frequently less spend, not more, because the constraint was never demand. The constraint was what the existing demand paid.

None of this says marketing never works. Constraints move. When capacity relaxes, an associate hired, intake systematized, hours freed, Acquisition can become the binding constraint, and where demand genuinely binds, intake systems and referral channels demonstrably move revenue. The discipline is one constraint at a time, named by the evidence, and the evidence on a fully booked solo rarely names demand.

There is a zone where a guardrail belongs, and it can be named without setting it. Spend that lands work whose contribution per owner hour sits below the firm’s current book is load, not growth. Where the line sits for a given firm depends on its fee structure, its collection reality, and the real cost of the owner’s hour, and none of that is answered here. Naming the zone is enough to change what the renewal signature means. The question it replaces is whether the firm is busy. The question it installs is whether the next dollar of demand clears the book it lands on.

The calendar will still be full next month either way. What it is full of is the finding.

Next week, Part 3: the ranking report, and what happened to the click it measures.

This issue is an economic diagnosis of a firm as a business. It is not legal advice and not the practice of law. It works from self-reported figures stated in ranges, is not a reconciliation, and is not accounting advice or a substitute for the firm’s accountant. It is not a valuation and not investment advice.

Sources. LEXGRO 2026 aggregation (74 percent low-ROI finding; spend bands), replicated by Amra and Elma and Seoprofy. Clio Legal Trends (attribution waste, 25 to 35 percent; spend bands). BestLawFirms / Best Lawyers survey, November 2025 (solo budget-cut rate, 24 percent; spend distribution). TMetric 2025 agency benchmarks and Swydo (utilization band, 65 to 80 percent; decline above 85 percent).

The first is an agency proposal. Growth engine, case studies, pipeline projections, a retainer number at the bottom. The second document the partner never sees. It exists anyway. It is the marketing industry’s survey of itself, and in it, 93% of the firms in the business of selling growth say their own growth engine is not strong enough. 7% call their pipeline strong.

The proposal and the survey describe the same companies.

This series reads the industry that sells growth by its own published numbers. Five parts, one per week. This one states the finding and spends one number from each of the four that follow. The demonstrations come week by week.

Same species, two panels

An agency is a professional services firm. A fixed labor base, sold by the hour or the retainer, with utilization, contribution, and a binding constraint. Same species as a law firm.

It runs on two instrument panels. The labor panel reads utilization, contribution per hour, and capacity. The demand panel reads leads, cost per lead, and pipeline velocity. The proposal on the partner’s desk is written entirely from the demand panel. The agency’s own survival is decided on the labor panel. The gap between the two panels is where this series lives.

One market-level pass makes the point. Worldwide ad spending grew 8.6% in 2025. Agency holding company revenue fell 1.2% in the same year. The growth was real. It went to the market and never reached the sellers.

Two shops, one claim

Take two shops at the industry’s average revenue, near $4.4 million, each with the word growth on its website. The first is a blended generalist: a 13% net margin and falling, utilization drifting under the industry’s own 70% floor, a pipeline fed by referrals it does not control. The second narrowed what it sells: 13% annual growth and a 30% net margin. Same headline revenue. Same year. Same market.

Neither grew through the product it sells. The winner grew by constraint. It cut its service list, and the margin followed the cut. Structure beat spend, and that pattern is what this series takes apart, on their books and on yours.

The figures here and throughout are industry composites, stated in ranges. The average shop and the narrowed shop are constructions from benchmark data. No firm in this series is a client, named or disguised.

Four forces, four essays

The justification problem. Marketing spend is treated as a growth decision. It behaves as a capacity decision wearing a growth costume. The buyers already sense the mismatch: 74% of law firm marketing budgets are reported going to low-ROI activities. Before the next retainer renews, there is a number to know first. That number is Part 2.

The ranking report. The monthly SEO report never changed. The click it measures did. On queries that now trigger an AI Overview, position one paid a 7.3% click rate in December 2023 and 1.6% by December 2025. The field name and the field’s meaning are no longer the same thing. Part 3 reads the instrument.

The menu and the mirror. Ask an agency what drives its own new business, and the industry’s survey has already answered. Then read the proposal it sent you. 20% of agencies say social media is not part of their own marketing at all, while selling it as a line item. The overlap between what they sell and what they use is close to zero. Part 4 runs the comparison line by line.

The agency reader. Written to the other side of the table. Agencies run traditional tactics on themselves as proof of product, and their books show the cost: 35% project margins collapsing to a 13% net, a $204,000 average pitch that loses two times in three. The sellers of growth need the same diagnosis, and they have never run it on themselves. Part 5 is theirs.

The question underneath

“Which agency will grow us” is the question the proposal is built to win. The question that governs the decision sits underneath it: what constraint binds this firm’s revenue system, and does an incremental marketing dollar clear the contribution test.

There is a zone where the answer is already visible. Spend that lands work whose contribution per owner hour sits below the firm’s current book adds load and calls it growth. Where the line sits inside that zone is a governance question, and this series will not set it. It will show you where the zone is.

The other panel

The proposal on the desk is honest. It is the panel the agency reads to itself, read out loud to you. Four essays from now, you will be able to read the other one.

If this read is worth your time, subscribe. The next four parts arrive weekly.

Subscribe now

The Billable Hour is an economic read of a firm as a business. It is not legal advice and is not the practice of law. It works from published industry data and self-reported figures and is not accounting advice or a substitute for the firm’s accountant. It is not a valuation. It is not investment advice.

Sources

• RSW/US, 2025 Survey Report “Rolling Toward 2026” (senior executives at marketing services and professional services firms, surveyed August 2025): 93% growth-engine finding; 7% strong pipelines.

• Forrester, Predictions 2026: Marketing Agencies, via Ritner Digital analysis: worldwide ad spend +8.6% in 2025 against agency holding company revenue of -1.2%.

• Promethean Research, 2026 State of Digital Services (119 agency leaders): average agency revenue $4.43M; average net margin 13%, down from 14%; narrowed-mix agencies at 13% growth and 30% net margins.

• SPI Research, 2025 Professional Services Benchmark: billable utilization 66.4% in 2025, first reading below the 70% floor, fourth consecutive year of decline.

• LEXGRO 2026 aggregation (replicated by Amra & Elma and Seoprofy): 74% of law firm marketing budgets reported going to low-ROI activities.

• Ahrefs, February 2026 CTR analysis: position-one click rate on AI-Overview-triggering queries, 7.3% (December 2023) to 1.6% (December 2025).

• SparkToro / Founder Focus, State of Digital Agencies 2025 (376 agency owners and consultants): 20% of agencies report social media is not part of their own marketing.

• ANA / 4A’s / Advertiser Perceptions, “Cost of the Pitch” (2023): $204,461 average non-incumbent pitch cost; two in three clients retained the incumbent.

The first subtraction

Write-offs for participating PPO providers average 30 to 40 percent of gross production (ADA Dental Fees Survey). A crown posted at $1,200 collects $720 to $840. The owner set the posted fee once, at credentialing. The payer has been setting the collected one ever since.

That arrangement is aging badly. Dental equipment and supply prices rose 5 percent through 2025 while reimbursement rates failed to keep up (ADA Health Policy Institute, Q4 2025). The ADA calls it the fiscal squeeze. Insurance is now the most cited concern among private-practice owners, named by 55 percent, and a third of the profession says it intends to drop at least one network. Legislatures noticed before most owners did: 37 dental insurance reform laws passed across 18 states in 2025, the largest single-year shift since the PPO model became dominant.

So the $965,660 becomes something near $600,000 to $680,000 in collections, and the terms were set by the other side of the contract.

The second subtraction

Overhead takes 60 to 65 percent of collections at a solo general practice, staff wages the largest line at 25 to 30 percent (ADA HPI and Dental Economics benchmark data). Rent, lab, supplies, and payroll do not adjust downward when a payer downcodes a claim or a patient’s history runs an appointment twenty minutes long. The practice absorbs those events at full cost and collects them at contract rates.

Run both subtractions and the ADA’s $215,320 is what remains. Twenty-two cents on the produced dollar.

The number still missing

Neither ADA figure has a denominator. Divide the $215,320 by the owner’s clinical hours and the result is contribution per owner hour: what an hour of the owner’s own time actually earned, after collection, after cost.

That number is invisible in the practice software and it is the only one that can rank the schedule. Two crowns with the same code, one in network at $780 and one out of network at $1,100, occupy identical lines on the production report and identical chair time. Per owner hour they are different procedures. A full book of the first can pay less than a lighter book of the second, and the production report will score the fuller book higher.

Where the owner runs hygienists or multiple operatories, the constrained resource can shift to chair hour or operatory day, and the same division runs on that denominator instead. Finding the binding one is the work.

The guardrail zone

Somewhere in every payer mix there is a line between a plan whose volume covers its own cost to serve and a plan whose adjustment exceeds the overhead its patients generate. A third of the profession intends to act on that line this year. Which network, and at what volume risk, is a threshold question, and thresholds are set from reconciled numbers, not from memory. This issue names the zone and stops.

The finding

A dental practice has never billed an hour. The production report is a fee total, silent on the owner hours underneath it, and everything the owner is paid from happens two subtractions and one division after the number it displays.

The fee schedule is memorized. The margin isn’t.

B.L. Sheets writes on the economics of founder-led firms, including revenue architecture, capital and transaction readiness, and pricing, margin, and retention. For law firm owners, the free Growth Intelligence Scorecard at growthprolegal.com reads the structure under your revenue in four minutes.

This is a general structural observation about the economics of owner-operated dental practices. It is not legal advice and not the practice of law. It is not accounting advice and not a substitute for the practice’s accountant. The figures are national averages and published benchmarks, stated as ranges where the source provides them, and drawn from no identifiable practice.

Sources: ADA Health Policy Institute, Survey of Dental Practice, 2025 (gross billings and net income). ADA Survey of Dental Fees (PPO write-off range). ADA Health Policy Institute, Q4 2025 State of the U.S. Dental Economy (reimbursement trends, owner concerns, network intentions). ADA News, January 2026 (2025 insurance reform legislation). ADA HPI and Dental Economics overhead benchmarks.