Incrementality testing that proves what works.

Incrementality testing measures the extra sales or leads your advertising actually causes. You compare a group that sees your ads with a control group that doesn’t, by people or by region, and the difference between them is the lift. It answers a different question from attribution: not which ad gets credit for a sale, but whether the sale would have happened without the ads.

A lift test, such as Meta Conversion Lift, splits people into test and control groups inside one platform. A geo test splits regions instead: ads run in some areas and not others, and you compare sales across them. Geo tests work for any channel, including ones that can’t split by person, such as YouTube, brand search, programmatic or out-of-home, and they measure total sales rather than platform-tracked conversions.

Usually two to six weeks of testing, plus time to design it beforehand and read it out afterwards. The length depends on how many sales you get and how big the effect is likely to be. We size every test before it starts, so it runs long enough to give a clear answer and no longer.

Enough for the difference between test and control to stand out from normal day-to-day variation. That depends on your sales volume more than your budget alone. Platform lift tests also have their own eligibility rules, and some need a Meta or Google account team to set up. We’ll tell you honestly whether your account is big enough before you commit.

Meta with Conversion Lift, Google Ads with conversion lift and campaign experiments where your account is eligible, and any channel with a geo test: Google search (including brand search), YouTube, TikTok, LinkedIn, programmatic, CTV and offline media. We also run holdout tests on retargeting and brand search, where platform numbers most often flatter.

Incremental ROAS (iROAS) is the revenue your ads actually caused, divided by what you spent. It’s usually lower than the ROAS a platform reports, because platform ROAS also counts people who would have bought anyway. It’s the better number for deciding where the next pound goes. Try the maths with our incrementality calculator.

Then you’ve found money to move. A low result rarely means switching a channel off overnight: we look at what the test covered, cut back gradually, and move budget to where the tests and the numbers say it will add more. A test that shows a channel works is just as useful, because it gives you the confidence to scale it.

Not quite. An A/B test compares two versions of something, such as two ads or two landing pages, and tells you which performs better. An incrementality test compares ads with no ads, and tells you whether the advertising is adding sales at all.

They work together. A marketing mix model estimates every channel’s contribution from your history and shows where it is least certain; a lift test then proves the biggest claim, and its result is fed back to calibrate the model. Smaller advertisers usually start with lift tests alone.

Yes. Incrementality tests can be designed and read out as a standalone project, alongside whoever manages your accounts. That gives you an independent answer about what’s working.

For management clients, lift tests are planned into the account as it grows. As a standalone project, each test is quoted up front once we’ve seen your channels and sales volume, covering design, set-up support and the read-out. Platform lift tests themselves have no extra fee from Meta or Google beyond your ad spend.