Performance reviews are usually discussed as an internal matter: development, compensation, promotion. For a firm doing institutional work, that framing is out of date. Clients have already written down what they expect from the people staffed on their matters, and they have put it in a document the firm signed.

Outside counsel guidelines are where a legal department converts its expectations on billing, staffing and conduct into a standard the firm can be held to. Three clauses appear in almost every set, and each one lands on the review system.

Three clauses, and what they ask of you

Three standard outside counsel guideline clauses and what each asks of a review system.
Typical clauseWhat it requires of the review system
No more than three attorneys billing a matter without prior written approvalFewer seats means each one has to be filled by someone the firm can defend. Reviews are how the firm knows who that is.
Staffing changes subject to prior client consent, and continuity expected for the life of the engagementEvery departure from a matter is a conversation with the client, not an internal HR event.
No payment for time spent bringing a new attorney up to speed when another rolls offAttrition mid-matter is written off as firm overhead. The cost of losing an associate is contractual, not notional.

The third clause is the one that changes the arithmetic. Zscaler's published guidelines state the firm will not be paid for training time when a new attorney is staffed to replace one who transferred off. HESAA's are the same. That is standard language, not an outlier.

So when a mid-level associate leaves a matter, the firm absorbs the replacement's ramp-up at its own expense, and asks the client's permission to do it. Whatever the review system does or fails to do about retention shows up there.

Where firms lose the thread

Most review instruments ask about the things a firm can count. Hours, originations, matters closed. Clients scrutinising staffing efficiency are asking a different question, and the 2026 Outside Counsel Benchmarking Report shows average team sizes in litigation and M&A falling as legal departments press harder on who is doing what.

A smaller team means less room to carry someone. It also means the difference between a well-supervised third-year and a poorly supervised one is visible to the client in a way it was not when six people billed the matter.

The tenure problem underneath

Our benchmark data puts indexed engagement at 82 in year one, 68 in year two, and 47 across years three and four. Years five and six recover to 55, and partners sit at 74.

The trough falls on the mid-level associate. That is the same person a client most wants continuity from: experienced enough to run the day-to-day, cheap enough to staff heavily, and holding the working knowledge of the matter. A firm with a year three and four engagement problem has a client continuity problem it has not named yet.

What to change in the instrument

Four adjustments are worth making, roughly in this order.

Score the behaviors the guidelines actually require. Responsiveness, handover quality, and whether work was pushed to the right level of seniority. These are assessable and they map directly onto clauses in the engagement.

Ask the associates, not only the partners. Supervision quality is the largest controllable input to whether a mid-level associate stays on a matter, and the only people who observe it are the ones being supervised.

Report by matter team where the population allows it. Firm-wide averages hide the group where continuity is breaking. Where a team is too small to report on safely, aggregate it rather than estimating.

Put review data next to client feedback. A relationship partner hearing about responsiveness from a client, and a practice group leader seeing responsiveness scored internally, are working on the same problem from two directions and usually do not know it.

What this is not

None of this means showing clients your review data. It stays internal, and it should.

It means accepting that the client already specified the standard, in writing, with money attached. A review system built around billable hours measures something the client explicitly said it is not buying.