Field Note · Measurement
Your calibration multiplier has an expiration date
A brand runs a lift study in Q3. The result: a 0.5x calibration multiplier. Six months later, they overhauled their creative strategy, changed attribution windows, and shifted budget across funnel stages. The calibration multiplier hasn't moved. They're making every budget decision on a number that no longer reflects reality.
Here's the backstory. Running a Conversion Lift study or geo-holdout test gives you the true incremental value of your ad spend. It's the most conservative read on the time-bound impact of your investment. That's the gold standard.
But the test ends. You go back to your primary attribution dashboard, where the numbers you see every day are attributed conversions, not incremental ones.
So you do the smart thing. You compare. The lift study says 100 incremental conversions. Your dashboard says 200 attributed. You calculate an incrementality adjustment factor: 0.5x. Now you can translate daily reporting into something closer to reality.
This is your calibration multiplier. And it is the most operationally important number in your measurement stack.
The problem: that calibration multiplier is not permanent.
It varies by funnel stage, campaign objective, region, optimization type and more. It can shift every time something changes in your strategy or in the attribution system. A platform update. A new attribution window. A change in how clicks or views are counted. Your own conversion API and pixel implementation. Any of these can move the denominator (attributed conversions) or the numerator (true incremental conversions).
I've watched this play out with brands at every stage of scale. A brand calculates their calibration multiplier and then coasts on it for 6 to 12 months. Meanwhile, they've made significant changes to their strategy; new funnel-focused campaigns, different attribution settings, new offers or creative. Their strategy has shifted. Their calibration multiplier hasn't. And they're making budget decisions on a number that no longer reflects what actually happened.
Here's the part that really hurts. When calibration drifts, you can't diagnose your own performance correctly. You're staring at a reporting dashboard asking "why is ROAS declining?". The team is checking ad delivery changes, funnel and conversion issues, signal gaps, audience saturation. But the real answer might be: ROAS isn't down. Your measurement shifted.
When you can't tell the difference between a performance problem and a measurement problem, every diagnostic question becomes a guessing game.
The fix isn't a one-time recalibration. It's treating calibration as an ongoing system, which is where this series goes next.