Planning assumptions to test with marketing agencies

For buyers reviewing marketing agencies in England, this guide sets out agency-side planning assumptions for 2027. Use them to discuss client budgets, AI-assisted work, scope and the triggers for reviewing an agency's commitments.

What to take away

The rate figures above are a dated snapshot, not a full sequence of Monetary Policy Committee decisions or a schedule for 2027. Check the latest Bank of England announcement before using them in a budget discussion.

Which economic signals should an agency monitor?

Start with the cost of money, because it sets the tone for every pitch. The Bank of England's Monetary Policy Committee held Bank Rate at 4% on 18 June 2026, having cut it to that level in May 2026, and the rate has been lowered again since. That is a materially different backdrop from the 2023 and 2024 peaks, but it is not cheap credit. The Committee announces decisions eight times a year, so an agency plan for 2027 will meet several rate decisions before it is a year old.

For an agency in England, the practical effect is that clients are willing to commit to annual plans again, provided the plan is defensible to a finance director. Marketing budgets are no longer being frozen mid-quarter as a default response to rate news. They are being rebuilt around fewer, larger programmes.

The Bank of England's own framing of the economic outlook is worth reading directly, because it shapes how cautious a client's board will be (Bank of England). If the central bank is signalling gradual easing, agencies can argue for multi-quarter work. If it is signalling caution, agencies should expect shorter commitments.

One practical habit helps here. Quote every proposal with a scenario table showing what the client gets at three spend levels, because procurement teams now ask for that anyway. A single headline number invites a yes or no answer. A tiered number invites a conversation about value.

The broader market picture — how demand is shifting between service lines, which sectors are hiring agencies and which are pausing — changes what you sell rather than how you plan. This guide stays with the assumptions to discuss with an agency: the signals it watches, how it would adapt scope and the triggers for a review.

Are client budgets actually recovering?

Not uniformly. The recovery is uneven by sector and by client size, and agencies that treat it as a single trend will misprice their work.

Consumer-facing clients in travel, hospitality and premium retail have loosened budgets faster than business-to-business clients in industrial and public-sector-adjacent markets. Public sector work in England remains subject to procurement cycles that are longer than the campaign windows they are meant to serve.

A useful discipline is to separate budget conversations into three buckets: money already committed, money contingent on a business case, and money that only appears if a competitor moves first. Most agency new-business effort is wasted on the third bucket.

Pricing follows the same logic. Retainer clients are asking for smaller core teams with named specialists brought in for defined periods. That suits agencies with strong freelance networks and punishes those carrying heavy permanent benches.

A freelance-heavy team brings tax rules into the plan. Where a specialist works through their own limited company, HMRC's guidance on off-payroll working (IR35) says that in most cases the client is responsible for determining the worker's employment status, while for a small client outside the public sector the worker's intermediary decides. An agency that engages contractors this way should know which side of that line it sits on before it prices a flexible team.

Industry agendas give a reasonable read on where client attention is going. The ISBA past events archive shows the topics brand-side marketers have been convening around, which is a decent proxy for what will appear in briefs (ISBA past events).

What does AI adoption mean for how agencies sell?

AI has stopped being a differentiator in pitch documents and started being an assumption in them. Clients now ask what the agency does with it, not whether it uses it.

The IPA industry AI resources are a sensible starting point for agencies that need to show clients a policy rather than a promise. They cover the practical questions clients raise: who reviews output, what data is used, and how the agency accounts for it commercially.

For personal-data work, UK data protection law comprises the UK General Data Protection Regulation (UK GDPR) and the Data Protection Act 2018. Everyone responsible for using personal data must follow the data protection principles unless an exemption applies.

The ICO's AI legal framework explains that AI using personal data falls within data protection legislation. Article 35 of the GDPR requires a Data Protection Impact Assessment (DPIA) where processing personal data, particularly using new technologies, is likely to result in a high risk to individuals. Ask the agency to explain whether that requirement applies to its proposed work, alongside its arrangements for data use and output review.

Three commercial models are emerging. Some agencies charge the same fee and absorb AI as a productivity gain. Some charge less and pass the saving on, which is a race to the bottom. The more durable model is to hold the fee and redirect the freed capacity into strategy, testing and measurement that clients could not previously afford.

Be explicit about which model you are running. Clients in England are increasingly willing to pay for judgement and unwilling to pay for production that a tool can do. Agencies that blur the two get squeezed.

AI also changes the skills mix. Junior production roles are narrowing while mid-level roles that combine editorial judgement with tool fluency are expanding. The hiring plan should follow the pricing model rather than lead it: recruit the mid-level judgement the retainer already pays for, and buy production in as the tooling improves.

Which trends will shape 2027 planning?

The most reliable planning assumption for 2027 is that clients will want fewer agencies on the roster but deeper relationships with the ones that remain. Consolidation is being driven by internal marketing operations teams that are tired of managing five suppliers for one campaign.

Second, measurement will keep moving up the agenda. Clients are asking for incrementality evidence, not platform-reported reach. That favours agencies that can design a test, run it cleanly and explain the result without dressing it up.

Third, the boundary between agency and in-house is blurring. Many clients in England now run a hybrid model with an in-house lead and an agency for specialist execution. Agencies that position themselves as partners to the in-house lead rather than replacements for it are winning more work.

Fourth, procurement is arriving earlier in the process. That means agencies need commercial language, not just creative language, in the first meeting.

Treat these four as working assumptions rather than forecasts. Test each one against the next client brief and drop the ones the briefs contradict: a trend that never reaches a brief is a trend the agency does not need to plan around.

Industry activity is a decent leading indicator here. The IAB UK events list shows how much of the sector's attention is going into measurement, retail media and AI governance (IAB UK events). When the same theme appears across multiple events, it tends to appear in client briefs within two quarters.

Which operating shifts should agencies consider?

Five shifts separate how many agencies operated in 2024 and 2025 from the position to plan for in 2027:

Read these shifts as a planning tool, not a prediction of every agency's experience. A small digital marketing company serving local clients in England will feel them later and more gently than a networked agency pitching national accounts.

The pricing row is the one that changes agency economics most. An outcome-linked fee sounds attractive until the measurement is weak. That is why the measurement work has to be done before the commercial model is sold, not after.

How can agencies package scope and outcomes?

Move from hours to scope. Clients do not want to buy time; they want to buy a defined change in a metric they already report on.

A workable structure has three parts. A capped core retainer covers the standing team and the recurring work. A project pot covers defined pieces with a stated outcome. A performance element pays out only against an agreed measure with a baseline and a review date.

This structure protects margin because the core retainer is predictable and the performance element is bounded. It also forces the client to agree the measure up front, which removes the most common cause of end-of-contract disputes.

Be careful with the performance element. If the metric is not something the agency can materially influence, the model becomes a discount with extra steps. Set the review date where the number has had a fair chance to move, rather than tying it to the campaign end date.

The evidence has to exist before the fee is agreed. Name the baseline, the measure and the reporting cadence in the proposal, so the performance element is defined by a number the client already reports on rather than one the agency introduces at the review.

What could go wrong?

Three scenarios deserve a place in any 2027 plan.

The first is a slower-than-expected easing of borrowing costs. If Bank Rate stays higher for longer, client boards delay discretionary spend and agencies face a second year of flat budgets. The mitigation is to keep fixed costs flexible and to avoid long property or headcount commitments.

The second is a client-side hiring push that brings more work in-house. If clients recruit senior marketing operations people, they will want to own strategy and buy execution. The mitigation is to sell into that model rather than resist it.

The third is AI-driven price deflation in production-heavy work. If clients believe production should cost less, agencies selling production will lose margin. The mitigation is to move the value proposition towards judgement, testing and integration.

None of the three needs a forecast to plan for. What matters is that the plan names the signal that would show which scenario is unfolding, and the person who watches for it.

What should trigger a plan review?

Set explicit triggers rather than reviewing strategy on a calendar. A calendar review happens whether or not anything has changed, which wastes senior time.

Four triggers cover most agencies:

When a trigger fires, review three things: the pricing model, the service mix and the skills plan. Do not review the positioning statement, because that is rarely the problem. Write the triggers down and name who monitors each one, or the exercise becomes an annual meeting with no decision attached.

For a buyer, use these triggers in the agency review rather than accepting them as a forecast. Ask the agency to show which assumptions apply to your brief, what it would change in the scope or team if they fail, and who can authorise a change.

Keep those decisions in the proposal alongside the baseline, review date and boundaries of any performance element. A useful agency response connects its assumptions to the work you are buying; a general account of industry trends does not.

For agencies operating across the UK, remember that some rules differ by nation. Employment rights rest mainly on the Employment Rights Act 1996, with ACAS guidance on procedure; public sector buying in England is now governed by the Procurement Act 2023, which came into force on 24 February 2025, while the Public Contracts Regulations 2015 continue to apply to procurements started under the old rules; and advertising claims are judged against the CAP Code, enforced by the Advertising Standards Authority. Procurement and certain advertising requirements can vary between England, Scotland, Wales and Northern Ireland, so check the position for each nation where you employ people or bid for public work.

Common questions

Is the marketing agencies outlook for England improving or worsening?

It is improving unevenly. Easing borrowing costs since 2024 have restored some client confidence, but budgets are being re-scoped rather than expanded, so agencies should plan for flat volume with higher expectations per pound spent.

How quickly is AI changing agency pricing?

Faster in production-heavy services than in strategy. Clients now expect AI-assisted delivery as standard, which squeezes rates for pure production. Agencies that redirect the productivity gain into testing and measurement are holding fees better.

Should agencies move to outcome-based fees in 2027?

Only where the outcome can be measured credibly. A capped retainer plus a bounded performance element works when there is a baseline and a review date. Without those, outcome pricing is a discount in disguise.

What is the single biggest risk to agency margin next year?

Unpaid senior time. Review cycles, procurement processes and compliance work are consuming more director hours than they did in 2024, and most retainer models do not recover that cost.

Do the same rules apply across the UK?

Not always. Employment rights, public procurement and some advertising requirements differ between England, Scotland, Wales and Northern Ireland — the Employment Rights Act 1996, the Procurement Act 2023 for public tenders in England and the CAP Code, enforced by the Advertising Standards Authority, are the main reference points — so agencies operating UK-wide should verify the position for each nation rather than assuming a single UK rule.