Incrementality for CMOs: how to answer "what did marketing actually do?"

Every CMO eventually gets the question from the CFO: "If we cut this budget by half, what would we actually lose?" Attribution dashboards can't answer it, because they measure which ads were present before a conversion, not which ads caused it. Incrementality is the discipline of answering the CFO honestly.

The idea in one sentence

Incrementality is the difference between what happened with your marketing and what would have happened without it.

You can't observe "without it" directly, so you create it. You withhold marketing from a randomly chosen group of people, places or periods, then compare. Everything else is detail.

Where to look first

Some channels are much more likely than others to take credit for sales that would have happened anyway. Test these first:

• Brand search. People searching your name were probably coming anyway. The question is how many would click your organic listing instead.
• Retargeting and re-engagement. Showing ads to users already on their way back can look brilliant in attribution and add very little.
• "Prospecting" that isn't. App-install campaigns that reach mostly existing or lapsed users, often through broad lookalikes.
• Affiliate and coupon partners that sit at the last step of the purchase.

The three tools, and when to use each

Tool	Best for	Watch out for
Platform lift studies (conversion lift, brand lift)	Quick reads on a single platform's campaigns	The platform designs and grades its own exam. Use it for direction, not as the final word.
Your own holdouts and geo tests	Settling specific debates (does retargeting work for us?) with your own data	Needs enough volume and clean randomisation. Gulf markets have relatively few large geos.
Media mix modelling (MMM)	Big budget allocation across channels, including offline, over quarters	Needs two-plus years of decent data. Calibrate it with your test results.

Open-source MMM tools such as Meta's Robyn and Google's Meridian have made modelling far more accessible. They are not magic, though. A model fed with messy spend data and no experiments will confidently tell you the wrong thing.

Geo testing in the Gulf: possible, with care

Geo tests switch marketing off (or on) in some regions and compare them with similar regions. In the US you might have hundreds of regions to choose from. In the Gulf you have fewer, and they differ a lot. Dubai is not Abu Dhabi, and Riyadh is not Jeddah. Practical tips:

• Use city or emirate-level splits within one country rather than comparing countries with each other.
• Run tests longer than you would in larger markets, and avoid Ramadan, Eid, White Friday and other big retail peaks.
• Prefer "turn down" tests (reduce spend in test regions) for expensive channels. They are cheaper and often more politically acceptable.
• Write the success metric and decision rule down before the test starts.

How to present it to the CFO

1. Lead with the decision: "We are moving X from retargeting to prospecting because a four-week holdout showed Y."
2. Show the counterfactual plainly: test group versus control, with a confidence range, not a single heroic number.
3. State what the test can't tell you.
4. Commit to a testing calendar, one meaningful test a month, so the conversation becomes routine rather than a crisis.

A starter programme for the next two quarters

• Month 1: brand search holdout in one market.
• Month 2: retargeting holdout (randomised user-level, if your platforms support it).
• Month 3: a geo test on your biggest prospecting channel.
• Months 4–6: stand up a simple MMM, calibrate it with the three results, and use it to set next year's budget split.

Bring your test design to a Gulf Growth Club session. Nothing helps a holdout more than five people trying to poke holes in it before you spend the money.