[ Incrementality testing & media mix modeling ]
Stop guessing with attribution. Start measuring.
Find out which marketing actually makes you money.
Multi-touch attribution gives credit to whatever ad was clicked last, not to what created the sale. RevClick replaces guesswork with media mix modeling (MMM) and incrementality testing, and measures every channel against the number that matters: profit.
- Cost per approval at Capital One, with guardrails set after a brand-keyword incrementality test
- -15%
- Brand ad spend as a share of revenue at Signal Relief
- -65%
- Total ad spend at Lignomat, with store revenue within 6%
- -74%
[ Why attribution falls short ]
Multi-touch attribution is costing you money
- 01
It credits the closer, not the cause
Retargeting and branded search claim sales that other channels created.
- 02
It can’t see everything
TV, YouTube views, podcasts and word of mouth leave no click, so they get no credit.
- 03
It breaks with privacy changes
iOS updates and cookie limits leave growing gaps in the data.
- 04
It rewards waste
Over-crediting bottom-of-funnel ads pays for customers who were going to buy anyway.
[ How it works ]
The RevClick standard: MMM plus incrementality
[ Media mix modeling ]
The big picture
MMM uses your historical spend, sales, seasonality and outside factors to estimate how much each channel contributes to revenue. It needs no cookies or user tracking, and it covers offline channels like TV.
[ Media mix modeling ]The big picture
MMM uses your historical spend, sales, seasonality and outside factors to estimate how much each channel contributes to revenue. It needs no cookies or user tracking, and it covers offline channels like TV.
[ Incrementality testing ]
The ground truth
We isolate a channel or campaign and run a controlled experiment: show ads to one group or region and hold them back from another. The difference in sales is that channel’s real value.
[ Incrementality testing ]The ground truth
We isolate a channel or campaign and run a controlled experiment: show ads to one group or region and hold them back from another. The difference in sales is that channel’s real value.
[ Calibration loop ]
A model that keeps getting more accurate
- Baseline: build the first model to see where spend looks efficient
- Hypothesis: find a channel where the model and the platform disagree, for example a platform claiming 10x ROAS where the model sees 2x
- Test: run a lift test to measure the real incremental return
- Calibrate: feed the result back into the model, then pick the next question
[ Calibration loop ]A model that keeps getting more accurate
- Baseline: build the first model to see where spend looks efficient
- Hypothesis: find a channel where the model and the platform disagree, for example a platform claiming 10x ROAS where the model sees 2x
- Test: run a lift test to measure the real incremental return
- Calibrate: feed the result back into the model, then pick the next question
[ Proof ]
Brand keywords are the usual first test
Paying for your own brand name often looks like your best campaign. At Capital One we tested the incremental value of branded keywords and set guardrails so that spend stayed profitable. At Signal Relief, separating brand from everything else cut brand spend from 2.6% to 0.9% of revenue.
[ Proof ]Brand keywords are the usual first test
Paying for your own brand name often looks like your best campaign. At Capital One we tested the incremental value of branded keywords and set guardrails so that spend stayed profitable. At Signal Relief, separating brand from everything else cut brand spend from 2.6% to 0.9% of revenue.
[ Case studies ]
Results from measuring what’s real
Capital One Traditional Agency
-15%
Cost per approval on low-volume keywords, with 10% more volume at the same unit economics
Read the case studySignal Relief AI Marketing Systems + Fractional CMO
1
Source of truth for every channel, with AI running analysis, paid search and agency checks
Read the case studyLignomat Traditional Agency
-93%
Cost per brand click, with ad spend down 74% and store revenue within 6%
Read the case studyFundbox Fractional CMO
-40%
Cost per application while ad spend scaled 300% and lifetime value per account rose 10%
Read the case study
[ FAQ ]
Common questions
How much data do I need for media mix modeling?
Generally a couple of years of weekly spend and sales data across channels gives the most reliable model. With less, we lean more on incrementality tests.
What is an incrementality test?
An experiment that compares sales where ads ran with sales where they didn’t, such as matched regions or a randomized holdout group. The difference is the sales the ads truly caused.
Will testing hurt my sales?
Tests are designed to be small and time-limited, usually holding back a modest share of spend or a few regions for a few weeks.
Does this replace my attribution tool?
No. Attribution is still useful for day-to-day optimization. MMM and testing tell you how much to trust it and how to set budgets.
[ Get started ]
Do you know which marketing spend is incremental?
Tell us your channels and budget. We’ll suggest the first test we’d run and what it could save you.
- A senior marketer reviews your request, not a sales rep
- You’ll hear back from us shortly
- No long-term contract required on most plans