The State of Legal Practice Management 2026
Five findings about how firms actually run their practices — gathered while building and deploying practice software, and written down with the method and the limitations attached, because an analysis you cannot interrogate is just an advertisement.
What this analysis is based on
- Structured conversations during pilot onboarding and demonstrations with firms of between one and forty fee-earners.
- Public product material from the four categories of tooling compared in this section.
- Our own build record — the numbered runbook of decisions, deployments and mistakes kept while building the product.
- Published practice guidance on client money, record-keeping and supervision, used to frame the confidentiality questions below.
What it is not
- Not a statistically representative survey. No random sample, no weighting. Treat the findings as patterns worth testing in your own firm, not as market share.
- Not vendor-funded research. Nobody paid for a favourable finding. We are the vendor, which is itself a limitation, stated plainly.
- No firm’s data was examined. Nothing in this analysis comes from reading anybody’s client files.
- No numbers invented. Where we have no data, we say “we do not have the data” rather than dressing an impression as a percentage.
Five things firms consistently recognise in the first ten minutes of a demo
Each one is a pattern we hear repeatedly — and each one is checkable against your own firm this week.
Most firms are running two systems of record
The practice software holds the matters. A spreadsheet holds the truth about deadlines, or fees, or who is doing what — because the software does not quite fit, and the spreadsheet is faster at answering the question the partner just asked.
The cost is not the duplicate typing. It is the contradiction: two answers to the same question, and no agreed way to decide which one is right.
The client portal is the least used feature — and the most requested
Firms ask for a portal, then leave it switched off, usually because it was never populated or because clients were not told about it. Meanwhile the same firms field daily calls asking for exactly the information a portal would have shown.
The pattern that works: open the portal on one matter with a willing client, then let the client’s own habit do the persuading.
Time capture fails at the point of recording, not the point of billing
Nobody argues about the invoice. The loss happens hours earlier: the timer that was never started, the note written after a long call, the “I will write it up on Friday”.
Which is why the useful question is not “does your system support time recording?” — every system does — but “how long does it take a fee-earner to record an entry, and what stops them skipping it?”
Confidentiality questions now come from clients and insurers, not only regulators
Firms report being asked — in client onboarding, in tenders and in professional-indemnity renewals — where their data sits and who else could theoretically reach it. A permission model is a harder answer to give than a structure.
The firms that handle this well have a one-page answer: where data lives, what is separated from what, who can access it, and what evidence exists. Written down, not improvised on a call.
Updates and exit are the least-tested procurement questions
Firms test features thoroughly and contracts lightly. The two questions that decide how the relationship feels in year three — how updates arrive, and what happens when you leave — are usually settled by a sentence in a proposal.
Both have concrete answers available: a release process you can describe, and an export you can see before you sign.
The gap, in one table
From the same conversations: what firms say matters when asked directly, and what the tools they shortlist tend to optimise for.
| What firms say they want | What most tools optimise for | What actually decides satisfaction |
|---|---|---|
| “Everything in one place” | Feature count and categories | Whether the file answers the partner’s question without a search |
| “Confidentiality we can explain” | Certifications and permission settings | The structure: what is separated from what, and whether it can be verified |
| “Get paid faster” | Invoicing and gateway integrations | Time captured when the work happened, and approved without a chase |
| “Less admin” | Dashboards and reports | Fewer re-typed details between enquiry, matter, document and invoice |
| “Clients kept informed” | A portal feature on the price list | Whether it is switched on, populated, and mentioned to the client |
| “No lock-in” | A documented way out and a data-export clause | Whether you have seen a complete export, and whether the software could run without the vendor |
A three-week evaluation that does not waste anyone’s month
Firms that evaluate slowly usually evaluate badly: the trial lapses, work gets busy, the decision is deferred. Three weeks, decided in advance, is enough.
Week one — the two problems
Name the two things that cost the firm the most time: deadlines, recovery, client updates, onboarding paperwork. Evaluate only against those. A demo that answers everything answers nothing.
Week two — put real work through it
Two open matters, one client, one invoice, one document request, done in the new system by the people who will use it daily — not by the partner who chose it.
Week two — test the uncomfortable parts
Ask for a real export. Ask what happens when a release goes wrong. Ask what happens to your firm when another firm has an incident. Ask to see the restore record.
Week three — decide, and write it down
One page: the two problems, the measured before and after, the cost, the exit route, and the decision. Firms that write this page rarely revisit the decision in a year.
What we would do differently with real survey data
This edition is a pattern analysis, not a survey. A proper one needs a random sample, a neutral questionnaire and results published whatever they say — including the findings that are unflattering to us.
- The survey — time to record an entry, status calls per week, hours reconciling client money, days to invoice after completion, and how long an export actually took.
- Independent analysis — we would rather a third party ran it and published the method.
- Published in full — including the parts that argue against buying our product.
Check the findings against your own firm
Two of the five will land immediately. We will show you what the other three look like once the numbers are visible — in your firm, not in a slide.