We agree one KPI from your own baseline data before anything starts - CAC, CPL, ROAS, conversion rate or revenue per visitor. Then we fix the whole path from click to revenue: ads, landing pages, tracking and the offer. If that number has not moved in 90 days and you followed the plan, our service fee comes back.
Revolut
Every company arrives describing a different problem. Underneath, the money is usually leaking in the same few places - and none of them is the ad account on its own.
Almost none of these is a single broken thing. It is a full-funnel problem wearing whichever symptom is loudest that month. Scaling a weak funnel does not create growth - it scales waste. So we find the actual constraint before we touch a budget.
A guarantee is only honest if we can reach every lever that moves the number. Most agencies can reach one of them. That is why their guarantee, where they offer one, is hedged into meaninglessness.
Account structure, targeting, budget allocation and wasted spend. A creative testing system with one hypothesis per test instead of six variables at once.
Above the fold, offer clarity, CTA, proof, form and checkout friction, objection handling, mobile. Prioritised by where users actually drop, not by what looks dated.
GA4, GTM, pixels and events audited and repaired. Duplicated conversions, messy UTMs and broken attribution resolved, so the baseline is real before anyone optimises against it.
Messaging, structure, proof blocks, forms. We rewrite and rebuild where that is the fastest path to the number, rather than filing recommendations someone else has to implement.
Often the real problem. The promise, the packaging, the risk reversal, and the match between what the ad said and what the page delivers. Good ads cannot rescue a weak offer.
For 90 days the entire system points at the single metric we agreed. Not at activity, not at a report, and not at whichever lever happens to be easiest to move.
Most guarantees in this industry are written to be unusable. Ours has conditions too, but you can read them before you buy rather than after a dispute.
If the agreed metric has not improved after 90 days, and you met the execution conditions below, we refund our service fee.
Why conditions exist at all: in past projects the single most common reason a metric did not move was that the recommendations were never implemented. We are not willing to guarantee an outcome we are not allowed to work on, and we would rather say that plainly than bury it in a clause you find later. Media spend is not refunded. It was never ours, it went to the platforms, and it bought you data either way.
This is a filter, not a formality. A refund promise made to a company with no traffic, no conversions and no history would be a sales tactic rather than a commitment.
Best fit sits higher up that range. The metric can be bad - the demand has to already exist.
If that is you we can still run the sprint, as a validation exercise without the guarantee, because there is no baseline to improve. We will say so before you pay, not after.
We look at the funnel, the accounts and the data you already have, then agree one primary KPI in writing. If your data cannot support a guarantee you hear that here, before money changes hands.
Tracking audited and fixed first, because everything after it is measured against this. Broken events, duplicated conversions and messy attribution get resolved before a single test runs.
Whichever lever is holding the number down gets attacked first, not whichever is easiest. Usually that is the offer or the page, not the ad account, whatever the last agency told you.
One hypothesis per test, enough data before a call is made, winners scaled properly and losers analysed rather than deleted. Creative insight feeds page insight and back again.
Measured the same way it was measured on day one. It either moved or it did not, and the guarantee resolves on that - not on a deck about how much work was done.
Our published work sits in high-intent verticals rather than in mainstream consumer brands, and we are not going to relabel it to look like something it is not. What transfers is the mechanic: buying the conversion behind the conversion instead of the cheapest one.
The cheap first conversion was worth nothing on its own. The budget was justified by the few that reached the second gate - the same arithmetic behind any trial, demo or basket.
Direct-to-consumer basket economics: the first order never paid for itself and the third one did. Creative burned in days, so the system had to out-produce the burn rate.
The funnel filled with people who would never qualify. The fix was making the creative do the disqualifying, before a sales team spent a week on someone who was never going to convert.
These are real campaigns in their true verticals. We have not yet published a mainstream consumer sprint under this offer - it is a new programme, deliberately capped at a small number of clients while we prove delivery.
Five fields. If you qualify we open the calendar straight away. If you do not, you will see that here rather than after a sales call.