How to validate lift from OOH / DOOH campaigns

My last post on OOH drew a lot of engagement.

One particular question from a commenter was “It is great to know that interaction effects in MMM can accurately measure impact of OOH. But how do we really validate that the OOH / DOOH campaigns worked?”

This is a great question and the answer to this is two fold – Geo Lift tests and Causal Geo Lift Tests.

📌 Geo Lift Tests

Let’s take an example

Suppose a brand launches DOOH only in selected cities/DMAs while keeping other similar geographies untouched.

You can now compare

▪️Search lift
▪️Conversion lift
▪️Store traffic
▪️Branded search volume
▪️Digital CTR improvements
▪️Sales uplift

between the Treatment Markets (with OOH/DOOH) vs Control Markets (without OOH/DOOH)

The key insight here is:

You are not merely measuring direct conversions from OOH.

You are measuring how the entire ecosystem behavior changed after OOH exposure.

This was step 1. Let’s look at a more robust causal way.

📌 Causal Geo Lift Tests – Difference in Differences (DiD)

In the last step, we kind of proved that the lift was because of the OOH/DOOH campaigns. But it wasn’t causal. To prove causality we need to control for a lot of factors.

Firstly we need a good counterfactual and second we should rule out natural evolution in a market.

Here is where DiD comes handy.

DiD asks:

“Did sales go up more in OOH exposed markets relative to non-exposed markets after adjusting for pre-existing trends?”

This is critical because many times markets naturally trend upward or downward.

Without DiD, organizations may falsely attribute macro movement to OOH.
The beauty of DiD is, it can validate not only direct lift but also halo effects.

For example:

– Did branded search increase disproportionately in exposed geographies?

– Did Meta CPA improve more in treated regions?

– Did retail footfall improve ?

📌 DOOH makes this even more measurable

Digital OOH introduces:

– Time based targeting
– Location granularity
– Dynamic creatives
– Programmatic serving
– Audience segmentation

All this makes application of causal geo experimentation much easier.

📌DiDective – our self serve causal geo test platform

DiDective is our self-serve Causal Experimentation platform built specifically for problems like validating OOH/DOOH.

It allows brands and analysts to run causal geo lift tests without needing a heavy econometrics background or setup.

But we also realized something important:

Running a causal experiment is only half the battle.
Interpreting it correctly is the harder part.

That is why DiDective comes equipped with our AI causal agent – DiDecto.

DiDecto can:

▪️Explain experiment results in plain business language
▪️Interpret treatment effects and confidence intervals
▪️Break down parallel trend assumptions

DiDecto also acts a causal inference learning repository. Ask it any causality questions 😎.

It is free to try for up to 5 experiments (link in comments).