The 2027 growth budget: what to test, what to reprice and what to cut first

A growth budget usually starts as last year's allocation adjusted for this year's targets. A more useful first step is to decide what will be cut, and in what order, if the year starts slowly.

Most businesses would cut by reported return on ad spend, removing the campaigns with the lowest platform return first. That order protects branded search and retargeting, the lines most likely to be credited with sales that would have happened anyway, and it removes prospecting, which looks weak on last-click credit because its effect arrives later and through other channels. A cut list built on reported return can remove the spend that was creating demand and keep the spend that was collecting it.

A better order uses evidence of cause. Spend that has been tested and shown to add little goes first. Untested spend aimed at people who already know the brand goes next, because its overlap with organic and direct demand is highest. Spend shown in a controlled test to add sales is protected longest. Building that list takes tests, and tests now fit much smaller budgets than they used to.

In November 2025 Google said it had lowered the minimum spend for incrementality experiments in Google Ads, so that a test that might once have cost upwards of $100,000 could be run for $5,000. Google has also made Meridian, its open-source marketing mix model, available to everyone since January 2025. Meridian can take the results of lift experiments as an input, so the model is calibrated against measured effects instead of resting on correlation alone.

For a business spending a few hundred thousand pounds a year on media, that changes the plan. One well-designed test a quarter on the largest uncertain line, read in the order book rather than in the platform, will settle more budget arguments over a year than an attribution dashboard. A geographic hold-out, where spend pauses in a set of comparable regions while it continues in others, is the most transparent design and the easiest to defend in front of a finance committee.

Three conditions make a test worth running. The outcome is net sales or qualified leads from the business's own records. The test runs long enough to cover the purchase cycle, which for a showroom, consultation or quote-led sale can be weeks. And the budget decision each possible result will trigger is written down before the test starts, so a disappointing result changes the budget instead of the test design.

Pricing changes usually pay back faster than media because they need no new demand. In the UK they now carry more regulatory weight.

Since 6 April 2025 the Digital Markets, Competition and Consumers Act has allowed the Competition and Markets Authority to enforce consumer law directly, with fines of up to 10% of global turnover, and it bans drip pricing, where mandatory charges appear late in the purchase. The CMA's first fine under the Act, a £4.2 million penalty, came in a drip pricing case that settled within five months.

Reference pricing is less settled, and that matters for anyone who runs seasonal promotions. CMA guidance written in 2024 for online mattress sellers expected one sale at the reference price for every two at the discounted price. On 30 July 2026 the High Court, in the CMA's case against Emma Sleep, declined to impose that fixed ratio and rejected the argument that low sales at a reference price are on their own enough to make it misleading, and the CMA then withdrew the guidance. The court still found infringements Emma had admitted, including a struck-through price for a product never sold at that price. A "was" price has to be genuine, offered for a meaningful period and supportable from records. The UK has no direct copy of the EU rule that a reference price must be the lowest charged in the previous 30 days, so a business trading in both markets runs two tests on every promotion.

For a 2027 budget, the promotional calendar is a compliance document as well as a trading plan. Keep a record of the price actually charged, and for how long, behind every reference price. Put mandatory delivery and service charges in the headline price. Measure price tests on contribution after discounts, returns and payment costs, by product and market, because a price change that lowers return rates can be worth more than its effect on conversion.

Retention plans for 2027 carry a new constraint for any business with UK consumer subscriptions. In August 2026 the government brought the start of the Act's subscription contracts regime forward to January 2027, from the spring 2027 date it had set in April. Businesses will have to give clearer information before sign-up, send reminders before renewal payments, offer a straightforward way to end the contract online and through customer service, and provide cooling-off periods at the start, after a free or discounted trial and on the renewal of certain longer contracts. The rules apply to contracts entered into after the regime starts. When the new date was announced, the secondary legislation and guidance had not yet been published.

Some retention figures depend on friction: a cancellation route that only works by phone, a renewal that arrives without notice, a trial that rolls into full price with little warning. Under the new rules those figures will fall, and a plan that carries this year's churn rate into 2027 will overstate revenue. The budget should model a higher churn rate for cohorts acquired after January and move retention spend toward the reasons customers stay, such as replenishment timed to actual use, service quality and loyalty terms worth keeping.

Retention is also where measurement pays most. Cohort revenue by acquisition month and channel, read from the order book, shows which sources bring customers who stay, and that figure should set the acquisition cost the business is prepared to pay.

A budget built this way is released in stages. A base allocation funds spend that has passed a test or that the business cannot do without, such as defending its own brand terms where a test has shown that it matters. A test allocation funds the quarter's experiments. A reserve is held back and released only when a test result clears a threshold agreed in advance.

That puts the growth budget on the same footing as other capital decisions. Each tranche carries an expected return, a measure read in the business's own records and a review date. Finance can see what was spent on learning and what was spent on proven demand, and the CMO can move money during the year without reopening the whole plan.

The same accountable team that sets up the tests should build what the results call for: the campaign structures, product feeds, price lists, checkout changes and retention flows, in accounts the client owns. A result that waits a quarter for a separate supplier to implement it has lost most of its value.

In November 2026, write the 2027 cut list in order of evidence and identify the two or three largest spend lines that have never been tested.

In December, design the first-quarter incrementality test on the largest of them, with the outcome measured in the order book and the budget decision for each result written down beforehand.

Before January 2027, review every UK consumer subscription flow against the new regime with legal advisers, and re-forecast churn for cohorts acquired after the start date.

By the end of the first quarter, audit the promotional calendar for reference prices and mandatory charges, and keep the record of prices actually charged behind every "was" price.

From April, release the reserve tranche only against test results. In June, rebuild the second-half plan from measured incremental contribution by channel.