Proven
A test that answered a clear question, at a budget and setup close to today's. We build on it.
The Lift Loop
How we run your ads, test after test.
The Lift Loop is our way of running paid media so it gets better over time. First we check your setup and connect your data, once. Then we go through the same six steps, again and again: plan the budget, run the ads, check them daily, test one thing, read the result, and decide what to keep, cut or scale. Each round starts from what the last test showed, so less of your money goes on guesses and more on what brings in customers. We call each round a loop.
Step by step
Set up once. Then six steps, every loop.
Click a step to see what happens there and which of our services do the work. Click a service to open its page.
Set up once, then six steps every loopBefore the test: reportedAfter the test: proven
↺ Results go back into the plan
Measurement & automationMedia strategy & buyingCreative & conversionDots under each step = services working there:leadssupports
Where we start
We start from what you already know.
Most accounts already have some tests behind them. In the audit we grade that evidence before anything new is tested: what still holds, what has expired, and what was never tested. Then the biggest gap goes first.
Expired
Once true, but the creative, budget, market or tracking has changed since. Goes back on the list.
A guess
Spend that has only ever been judged by the platform's own numbers. The biggest of these is tested first.
What changes over time
Each test changes where your money goes.
We run your campaigns and test what they add. Then we turn the result into a budget decision: back what works, cut what doesn't, and fund the next question. How many tests your budget allows, and how sharp the answers are, depends on your sales volume.
One €100,000 budget. Three decisions.
Illustrative. Total spend stays the same.Backed by a testBeing testedStill a guess
Paid social earns its place.
The €10k test adds sales at an acceptable cost. Keep it, and move the next €10k of guessed spend into a test.
Brand search needs less.
A test shows most of those sales would have come anyway. Cut €15k and move it to paid social, which a test already backs.
Video needs a clearer answer.
The result is inconclusive. Revise the test and keep its €10k in testing. The backed share stays at 45%, because nothing was proven.
Categories describe the evidence behind the spend, not guaranteed results. Old results are re-tested as conditions change.
How we test, honestly
The questions experienced buyers ask.
Testing has limits. These are the four we get asked about most, and how we handle them.
What if my budget is small?
A test needs enough sales to read, not enough euros. So we test the big questions first, where the honest answer might be "this adds almost nothing": brand search, retargeting. We split by region rather than by person, we read past pauses as free tests, and we run longer rather than bigger. What can't be tested gets labelled as a bet, not dressed up as a result.
What if one channel helps another?
Often it does: video makes people search your brand, and the sale lands on search. So we never measure a channel by its own numbers. We measure total sales in your own data, in the regions where the channel ran against the regions where it didn't. A channel that makes other channels work gets the credit, wherever the sale arrives.
What about things nobody controls?
A promotion, a PR story, a competitor's launch, another team's campaign. A test compares two groups at the same time, so these hit both groups and cancel out. What's left is what the ads did. If something only touched one group, we say so and the test doesn't count.
Can everything be proven?
No. Some lines are too small to test, results expire as creative and markets change, and new channels keep arriving. The aim is not 100%. It is that every euro is either proven, being tested, or a bet you chose on purpose, and you know which is which.
Every engagement starts the same way
Start with the audit.
Two weeks in your data, money back if it isn't worth it. If it's useful, the Lift Loop is what the ongoing work looks like.
Start with the auditThree questions on the home page, then we come back to you personally.