Outgrace, home
OperationsFramework

A framework for measuring service quality in professional firms

A five-pillar method for setting service measures on client work, tracking them and tying the results to renewal decisions. Built on editorial judgement; the research base is unverified.

Most professional firms measure their work by hours billed and invoices paid. Neither tells a client whether the service was good. A client can pay on time for years and still leave at renewal because the lead partner took too long to answer a question, or because the final report arrived in a form the client's own team could not use. The firm saw a healthy account. The client saw a slow one.

The gap is rarely a shortage of data. The firm knows how many hours went into a project. It seldom knows which promise to the client that project was meant to keep, or whether the promise was kept. This framework is written to close that gap. It suits a firm with a managing partner who wants a small set of numbers, a finance or operations lead who can produce them, and a client book that renews on a fixed cycle.

A word on the evidence. We could not verify a research base for this framework. The sources available to us do not cover service levels in client work, professional services management or the link between service measures and renewal decisions. The GOV.UK Service Standard page describes its purpose as helping teams create and run public services, but its extract lists no measures, so we have borrowed none. What follows is editorial judgement. Treat it as a method to test in your own client book, not as a finding.

Before you start

Three things must be true before any measure is worth tracking. First, the firm must agree on what it sells. If the partners describe a client engagement differently, the measures will reflect whichever description is most flattering. Second, someone must own the data. A measure that belongs to everyone belongs to no one. Third, the firm must be willing to hear bad numbers. If a breach is punished, the account lead will stop recording breaches, and the framework will report only good news.

Chapter one: Pillar one: write the promise for each engagement

Every client engagement needs a short written promise in working terms. It should say what will be delivered, in what form, by when, who answers questions and how quickly. Most firms hold some of this in a statement of work, but that document is written in the language of liability and is read by lawyers. The promise an account lead uses on a Monday morning should be a different document, one page or less, and it should sit in the engagement file where the team can see it.

The test is simple. Could the account lead read the promise aloud to the client, and could the client say that it describes what they bought? If the client would hesitate, the promise is too vague to measure. If the account lead would hesitate, it is too detailed to remember. Keep rewriting until both can say yes. Write one promise per engagement, not one per firm. A firm-wide standard hides the differences that clients notice first, and those differences are where renewals are lost.

Chapter two: Pillar two: choose measures the client would notice

Pick four to six measures for each client, and no more. Four families cover most engagements. The first is how quickly the firm responds to substantive questions. The second is how often deliverables are accepted first time. The third is whether milestones land on the agreed date. The fourth is whether cost stays within the agreed estimate. Each family can be measured with a count or a date. Choose the version the account lead can record without help from the central team.

Measures should differ by type of work. A compliance return has a fixed deadline and a sharp accuracy test. A strategy project has fewer fixed points and a looser idea of what good looks like, so its measures lean on responsiveness and on whether agreed decisions were taken on time. Write the four families once for each type of engagement, then adapt them to the client. Do not invent a new set for every account.

The test is whether the client would notice if the measure slipped. Utilisation, realisation and staff turnover matter to the firm, and they belong in the firm's own management reports. They rarely belong in a client review. A measure that only the firm cares about invites the wrong conversation, and the client quickly learns to ignore the whole set. If a measure does not change how the client feels about the work, drop it.

Chapter three: Pillar three: set a threshold and name an owner

A measure without a threshold is a chart, not a control. For each measure, write the line below which the firm has failed the promise. A first reply within one working day is a sensible starting point for many advisory engagements, though the right figure depends on the client and the work. Agree the threshold with the client where you can. Where you cannot, say so in the promise, so that the client knows the number was set by the firm and can challenge it.

Name one owner for each measure, usually the engagement lead, and give that person the authority to act on a breach. The test is practical. Ask the owner what they did the last time the measure breached. If the answer is nothing, the threshold exists only on paper. Owners change, so record the history of breaches in the engagement file. A new owner should be able to read, in ten minutes, what has gone wrong and what was done about it.

Chapter four: Pillar four: record the measure where the work happens

The measure should be captured when the event happens, not reconstructed at quarter end. A survey sent months after a project closes tells you how the client remembers the work. A log entry made the same week tells you what actually happened. A shared spreadsheet or the project tool your team already uses is enough to start. Do not wait for a new system before you begin. The discipline of recording matters more than the software that holds the record.

The test is the time between the event and the record. If the gap is longer than a week, the measure will be unreliable, and the team will know it. Ask the account lead to show you the last three entries and the dates they were made. If the entries cluster at the end of the month, the process is being run for the report rather than for the client. Fix that before you trust any trend line drawn from the data.

Chapter five: Pillar five: bring the measures into the renewal conversation

The renewal decision should draw on the same numbers the client has seen all year. Put the measures into the client review, held well before the renewal date, so the client can respond to a trend rather than a verdict. Show three things: the measures against the promise, the breaches and what changed as a result, and the open issues the firm is working on. Keep the pack short enough that the client reads all of it.

The test is whether the client would recognise every number in the review as one they discussed earlier. If a figure appears for the first time at renewal, the process has failed. Decide in advance what a breach trend will trigger: a change of team, a revised scope, a different price or, in some cases, a decision to stop working together. Write that rule down before the renewal conversation starts. A firm that decides after the numbers arrive will be tempted to read them so that they fit the answer it already prefers.

When the trend is bad, say so first. A client who sees a breach raised by the firm, with a plan attached, is more likely to stay than one who finds the same breach in a renewal pack. The conversation is not a defence. It is a record of what the firm noticed, what it changed and what it will watch next quarter. Clients tend to judge a firm by how it handles a bad quarter more than by how it performs in a good one.

Seven questions to test the framework

Answer each question for every live client engagement. Any no is a gap to close before the next client review.

  • Does each live engagement have a written service promise in working terms the client would
  • Are no more than six measures tracked for each client?
  • Would the client notice if each measure slipped?
  • Does every measure have a written threshold and a named owner?
  • Is each measure recorded within a week of the event it describes?
  • Will the client see every measure in a review held before renewal?
  • Is there a written rule, agreed before renewal talks, for what a breach trend will trigger

Outgrace checklist

Where it breaks

The framework fails in four predictable ways. Measures become a league table for account leads, who then manage the numbers rather than the client, closing tickets early or reclassifying requests. Thresholds are set in the firm's offices and never tested with the client, so the client judges the firm against a line it never accepted. Owners change and nobody inherits the history of breaches. The most damaging failure is the one described earlier: the numbers are assembled after a decision has already been made, to support it. Clients tend to notice that within a single renewal cycle, and the trust it costs is hard to rebuild.

The Service Standard helps teams to create and run great public services.

GOV.UK, Service Standard page, gov.uk

The GOV.UK page states a purpose, not a measure. Its extract lists no criteria for judging service, so the standard for a professional firm has to be set by the firm and agreed with its clients. That is harder than copying a published list, and it is the only version of the standard that a client will hold you to.

How to start on Monday. Choose three live client engagements: one large, one medium and one due to renew within the year. Write a one-page promise for each. Pick four measures per engagement, name an owner for each, and set a threshold you are prepared to defend to the client. Start logging the same week. Bring the first three months of numbers to the next client review. Before the renewal date, write down what you will do if a line is breached, and only then look at the figures.

Newsletter

The Briefing

One email a week: the leaders we profiled, the ideas worth stealing, and nothing else.

Weekly, free, one click to leave. Read our privacy policy.