Get your firm's free AI report card
BLOG POST · SOLO & SMALL FIRM PRACTICE

Your Firm Measures Hours. Clients Measure Something Else.

TL;DR
  • Hours, utilization and realization are lagging numbers that report what already happened, not where your practice is headed.
  • Pick a few client-facing measures, like response time, quote-to-engagement and repeat or referral work, and track them in a spreadsheet.
  • Measuring AI by hours saved rewards cognitive surrender; track whether you still review the AI drafts you sign.

Ask a law firm how it's doing and you'll usually hear about hours: hours billed, hours collected, and lately, hours saved by AI. Ask its clients and you'll hear something else entirely. An SPU course spent two lessons on that gap, and AI has made it a lot wider.

The course was The Solo Practice of the Future, and most of it was about value pricing. In 2013 it added a two-part written lesson, "Measure What Matters," that asked a question most firms still skip: are you measuring the things that predict your success, or just the things that are easy to count?

The scorecard most firms still use

The standard law-firm metrics all come back to the hour. Utilization is the share of the working day that turns into billable time. Realization is how much of that time actually gets invoiced. Collection is how much of the invoiced time gets paid.

The numbers aren't flattering. Clio's Legal Trends data puts the average law-firm utilization rate for 2025 just under 40%, roughly three billable hours in an eight-hour day, and lower still for solos and small firms (Clio). Firms respond by trying to push that number up, which means more time capture and more timesheets. If you've read SPU's older posts on KPIs for a successful law firm, you know these numbers have their uses. Instructor Ronald J. Baker's point was that they can't tell you where you're headed:

"Accounting reports are, by their nature, lagging indicators—like timing your cookies with your smoke alarm."

By the time a lost client shows up in your realization rate, the client is gone. The lesson wanted early warnings instead, and was blunt about the risk of only counting what's countable:

"If we only focus on what we can measure, we will become prisoners of our past, because it would be impossible to create a future that would be different than an extrapolated past."

How firms are measuring AI: in hours, again

Now look at how the profession talks about AI's payoff. Thomson Reuters' 2025 Future of Professionals survey found respondents expect professionals using AI to save five hours a week within the next year, up from a four-hour prediction in 2024 (Thomson Reuters; Attorney at Work). In the same survey, 53% of professionals said their organizations were already seeing a return on AI, but only 22% said their organization had a visible, defined AI strategy (Thomson Reuters; Lawyers Weekly).

Two things stand out. First, the headline measure of AI is hours saved, the same unit as the old scorecard. Second, the five hours is a prediction, not a measurement. Most firms don't yet have a theory of what AI should change, so they count time.

For a firm that bills by the hour, "hours saved" is a strange thing to celebrate. Every hour AI saves on billable work is an hour you no longer bill, which is the problem we cover in AI makes you faster, the billable hour makes that a pay cut. Hours saved only becomes good news when you decide what those hours are for.

Performance indicators vs. predictive indicators

The course's answer in Part II was to separate two kinds of numbers. A key performance indicator is a measurement, full stop. A key predictive indicator is different:

"A predictive indicator, by contrast, is a measurement supported by a theory, which can be tested and refined, in order to explain, prescribe, or predict behavior."

The lesson's example was Continental Airlines under Gordon Bethune in the mid-1990s. The airline had been obsessed with cutting cost per seat mile and ranked at the bottom on the measures customers noticed. Bethune focused the company on three: on-time arrival, lost luggage and customer complaints. None of them appears on a financial statement. All of them, the lesson argued, predicted whether customers would come back. The reason they worked:

"What makes the three indicators just listed leading is that they measure success the same way the customer does."

No client measures your firm by utilization. Clients measure whether you called back, whether the price matched the quote, whether the problem got solved, and whether they'd send a friend to you. Those are the leading indicators. Hours are what's left on the books afterward.

Two kinds of numbers

LaggingThe hours scorecardUtilization, realization, collection, AI hours saved"Like timing your cookies with your smoke alarm"
LeadingWhat clients measureCalled back? Price matched the quote? Problem solved? Would they send a friend?Measures success the way the client does
The smoke-alarm line is from the course lesson "Measure What Matters." By the time a lost client shows up in the realization rate, the client is gone.

That's the core of what the course taught on measurement. The lessons went further, into how theories get built and tested, a thought experiment about which indicators an airline CEO should watch every day, and why a credit score is a better model for a metric than a raw count. SPU's guest lecture with Jared Correia on the magic of key performance indicators is a practical companion if you want to set up tracking.

What not to measure: how much of the thinking you handed off

A theory of what AI should change in a law practice needs one more piece: where AI belongs and where it doesn't. John O'Connor, founder of Associate and SPU's Professor of AI Fluency, is direct about the risk:

"The lethal AI habit for lawyers is cognitive surrender of work product."

Cognitive surrender is letting the tool do the thinking and signing what comes out. It's why SPU's guest lectures on the ethics of using AI for lawyers spend so much time on verification. If your only AI metric is hours saved on legal work, you're rewarding exactly this. The fastest way to save hours on a brief is to stop reading it.

O'Connor's alternative is to point AI somewhere else first:

"Use AI to delegate everything around the practice, the business side, use AI sparingly where it is corrosive to focus and attention span."

The business side is intake admin, scheduling, billing follow-up, marketing operations and documents about the practice rather than for a client. It's the work solos have always struggled to hand off, as SPU's posts on how to dabble in delegating and on managing the business side of a solo practice make clear. AI is a much cheaper place to start delegating than a hire, and the risk to the client is lower.

A starting theory for a solo in 2026

The course's method is to state a theory, pick measures that would show if it's wrong, and adjust. Here's one theory a solo could test. This is our suggestion, not something from the course: if AI takes over the business side, I'll spend more of my attention on clients and on judgment, and clients will notice.

Measures that would test it:

  • Response time. How long from a new inquiry to a real reply, and from a client question to an answer.
  • Quote-to-engagement. How many prospects who get a price go on to hire you.
  • Scope surprises. How often a matter needs a change order or a fee conversation nobody expected.
  • Repeat and referral work. How many new matters come from past clients or people they sent.
  • Review time on AI drafts. Whether you're still reading what you sign. If this number trends toward zero, that's the cognitive-surrender warning light.

None of these is about the hour. Most are things you can track in a spreadsheet in an afternoon. And if the theory is wrong, the numbers will tell you, which is the whole point.

A theory a solo could test

1. State a theoryAI takes the business side; clients notice the attention
2. Pick measuresResponse time Quote-to-engagement Scope surprises Repeat and referral work Review time on AI drafts
3. AdjustIf the theory is wrong, the numbers will tell you
The method (state a theory, pick measures, adjust) is the course's. The theory and the measures are SPU's suggestion, not something from the course.

The lesson closed Part II with a question that fits a practice deciding what to do with AI:

"…if we get what we measure, isn't it time to start measuring what we want to become?"

For the other half of the argument, why the lawyer's judgment and sign-off are the product once AI drafts and researches, read lawyers are knowledge workers, and AI just proved it. More practice management pieces are in our solo and small firm practice archive. Your website is one of the few places you can watch a leading indicator directly: who inquires, and what they read before they do. SPU offers a free review of your firm's website.

Compare notes with other solos

Picking the right numbers is easier when you can see what other firms track. SPU membership gets you the community of solo and small-firm lawyers, monthly strategy sessions, and the SPU course library.

THE WEEKLY SPU BRIEF

Get one practical lesson for your firm every week.

Free, from practicing lawyers. Every issue includes a lesson from the archive. Unsubscribe anytime.

ABOUT THE AUTHOR
The SPU Team

The SPU Team writes and edits Solo Practice University's articles. Susan Cartier Liebel founded SPU in 2009 to teach lawyers the practical side of building a practice, the part law school leaves out, with courses taught by practicing lawyers. Today SPU is a community for solo and small-firm lawyers, with an AI Fluency program for firms putting AI to work. Nothing here is legal advice: check your own jurisdiction's rules.

About Solo Practice University →