How to Choose the Right Incrementality Test for Your Ad Spend Level

This is the final piece in my three-part series on Stella's incrementality report. If you haven't read the first two parts yet, go check those out first.

Today I want to talk about something practical: if you're going to run your own incrementality test, what should you actually be measuring?

This is where a lot of brands get it wrong. They try to run tests that are too granular for their spend level, and they end up with results that aren't statistically significant.

Stella has a really nice breakdown here that guides you on the type of test you should run based on your monthly channel-level spend.

👉🏼 [YouTube Explainer] - Part 3: Stella's Incrementality Test Structure

The Spend-Based Testing Framework

Your monthly ad spend determines what kind of incrementality tests will actually give you meaningful results. Here's how it breaks down:

Lower Spend: Focus on Channel-Level Holdouts Only

If you're spending less than $100k per month on a given channel, your incrementality tests should be focused primarily on channel-level holdouts.

What does this mean?

This would be a geo holdout for only your Meta ads, or a geo holdout for only your TikTok ads—rather than any sort of tactical-level incrementality test.

At this spend level, you don't have enough volume to get statistical significance on more granular tests. Stick to the basics and test whether the entire channel is incremental.

Mid-Tier Spend: Channel Plus Tactic Level Tests

Once you're spending between $100k and $150k per month, you can move to channel plus tactic level testing.

This is where things get more interesting.

If you wanted to test the incrementality of Performance Max versus standard Shopping campaigns, or Meta Advantage+ Shopping campaigns specifically, you would actually need to have a spend level closer to six figures a month for this to give you statistical significance.

This is the tier where most serious DTC brands should be operating. You're not just asking "is Meta incremental?" but "is this specific campaign type within Meta incremental?"

Higher Spend: Campaign-Level Testing

Between $150k to $200k monthly spend, you can get down to campaign-level testing.

We do very little of this, but it can be interesting if you have a specific Performance Max campaign, a specific Meta ASC campaign, or a Connected TV campaign that you want to run a holdout test on.

That's something you can do, but you see the requirements of the spend to do such a thing.

Enterprise Spend: Advanced Testing Options

Once you get above $200k monthly spend, you have all sorts of additional types of incrementality tests available to you.

But for the most part, most brands are going to fall in these first two buckets—focusing on channel-level and tactic-level holdouts.

Why This Framework Matters

Here's the thing: running incrementality tests costs money. You're literally turning off spend in certain geos or to certain audiences to see what happens.

If you don't have enough volume, you're just burning money on a test that won't give you actionable insights.

The guidance from Stella here is incredibly valuable because it prevents you from wasting budget on tests that are too granular for your spend level.

Channel-Level Tests Tell You the Most Important Thing

Even if you're at the lower spend tier and can only run channel-level holdouts, that's still giving you a wealth of information.

You're answering the fundamental question: "Is this entire channel actually driving incremental sales, or am I just getting credit for sales that would have happened anyway?"

That's massive. That alone can reshape your entire media strategy.

Tactic-Level Tests Optimize Within Channels

Once you can afford tactic-level tests, you're moving from "should we use this channel?" to "how should we use this channel?"

Should you run Performance Max or standard Shopping? Should you use Advantage+ Shopping or manual campaigns? These are questions that can only be answered with proper incrementality testing—and only if you have the spend to support statistically significant results.

The Bottom Line

Don't try to run tests that are too sophisticated for your spend level.

Start with channel-level holdouts. Prove incrementality at the macro level first. Then, as you scale spend, you can get more granular with tactic-level and campaign-level tests.

But if you're spending $50k a month on Meta and trying to test individual campaign performance, you're wasting your time and money. The sample size isn't there.

Focus on the tests that will actually give you actionable insights at your current scale.

Want help designing incrementality tests for your brand? We work with clients across different spend levels to implement the right testing framework. Send us a message to discuss your specific situation.