The Client Promise Ledger: Writing Agency OKRs That Are Honest to Both Sides

Why client-facing OKRs need a different standard

Internal OKRs and client-facing OKRs may look similar on a planning page, but they carry different weight. An internal objective can be deliberately ambitious, reviewed in private, and allowed to fall short if the learning justifies the attempt. A client-facing objective is a promise. It sits close to the statement of work, the retainer agreement, and the relationship itself. For UK agencies and professional-services firms, that distinction is not academic. It shapes trust, scope control, and the credibility of every progress review.

In 2026, many agency teams are under pressure to show measurable client value without turning every project into a contractual performance guarantee. The answer is not to abandon client-facing OKRs. It is to write fewer of them, make them more honest, and keep the language of aspiration away from the language of commitment.

The distinction that matters: stretch versus promise

A stretch goal is an internal bet. It says: we believe this is worth attempting, we will learn from the attempt, and we will not treat a near miss as a breach of faith. A client promise is different. It says: this is the change we are prepared to be held to, on evidence both sides accept, with dependencies named and reviewed.

Agencies get into difficulty when they present stretch goals as promises. A bold internal target such as doubling a client's conversion rate may be a useful planning device inside the agency. Presented externally without qualification, it becomes a projection the agency cannot fully control. The client hears a commitment. The agency hears an ambition. That gap is where disputes begin.

The practical rule is simple: keep stretch goals internal, and keep client-facing OKRs narrow enough to survive a difficult quarter.

A two-column test for client-facing objectives

Before any objective reaches a client document, place it through a two-column test. In the left column, put candidates that belong in a client-facing OKR. In the right column, put candidates that should remain internal, or should be rewritten before they are shown.

Left column: promise candidates

An objective belongs on the left if it passes four tests:

• It describes an outcome the client would recognise and value, not an activity the agency can bill.
• It can be evidenced with data or artefacts both sides accept.
• It sits within the agency's reasonable influence, with client dependencies named rather than assumed.
• The agency could repeat the same commitment to another client without rewriting the contract each time.

Right column: projection candidates

An objective belongs on the right if it fails any of those tests. Typical right-column items include:

• Deliverable counts, such as publishing twelve articles or running six workshops.
• Speculative gains, such as doubling revenue or transforming a market position.
• Outcomes that depend heavily on client behaviour but are written as if the agency controls them.
• Internal stretch targets that would embarrass the team if published as a promise.

The test is deliberately blunt. It forces a useful conversation before the client sees the document. If a line cannot survive the left column, it is not a client promise. It is a hope, an activity, or an internal challenge.

Keep the ledger small

A client-facing OKR is not a substitute for a statement of work. The SOW defines scope, responsibilities, fees, and change control. The client OKR defines the change the engagement is meant to produce. Mixing the two produces bloated documents that nobody reviews after month one.

A workable pattern is one client-facing objective per major engagement, supported by two or three key results. A second objective may be justified for a long retainer, but only if the two are genuinely distinct. The moment a client OKR starts to read like a project plan, it has stopped doing its job.

This is where a promise ledger helps. The agency keeps a short record of what has been promised externally, what evidence will be used, and which dependencies sit with the client. Internal stretch goals stay in a separate ledger. Both can be reviewed, but they are not presented as the same kind of commitment.

Case study: Northbank Studio

Northbank Studio is a 24-person digital agency in Bristol. In late 2026, it was preparing a renewal for a regional independent school group. The previous statement of work had grown to nineteen deliverables, including a website audit, a CRM review, a content calendar, an enquiry-form redesign, staff training, and a monthly reporting pack. The client was not short of activity. It was short of a clear promise.

The agency replaced the long deliverable list with a single client-facing OKR: by 30 June 2027, raise the share of online admissions enquiries that reach a qualified conversation within one working day from 45 per cent to 75 per cent, without increasing admissions headcount. The supporting key results were specific but not overreaching: reduce the enquiry form from eleven fields to six, introduce automatic acknowledgement and routing, train two admissions staff, and publish a weekly evidence dashboard.

The agency also kept an internal stretch objective: pilot an AI enquiry-summarisation tool by September 2027. That objective was useful inside the team, but it was never presented as a client promise. The client dependencies were named in the agreement: timely access to the admissions inbox, two staff available for training, and a named client lead for sign-off.

Ten weeks into the engagement, the share of enquiries reaching a qualified conversation within one working day had reached 61 per cent, and the median response time had fallen from 1.8 days to 1.1 days. The agency did not claim success early. It used the review to surface two client-side delays and to adjust the evidence pack. The renewal conversation moved from a debate about deliverables to a discussion about whether the promised outcome was still the right one. That is the practical value of a client promise ledger: it makes progress legible without turning ambition into overcommitment.

What belongs in the agreement

A client-facing OKR should be supported by four things in the agreement or its appendices: the evidence source, the review cadence, the named dependencies, and the change-control route. The objective should not read as a legal guarantee of a business result. It should read as a commitment to pursue a defined outcome, measure it honestly, and review it with the client at agreed points.

That language matters in UK professional services, where retainer clients and public-sector buyers increasingly expect clear outcomes but also understand shared dependency. The agency that says plainly what it can influence, what it cannot, and how both sides will know whether the promise is being kept is usually the safer partner.

The discipline of the ledger

The client promise ledger is not a bureaucratic extra document. It is a discipline for keeping two kinds of ambition apart. Internal stretch goals keep the team hungry. Client-facing promises keep the relationship honest. When an agency confuses the two, it overpromises in public and underlearns in private.

A small set of evidence-led client OKRs, reviewed against named dependencies, will usually do more for trust than a long statement of work full of activity. The test is whether the client can repeat the promise back to you without exaggeration. If they can, the ledger is working.