Scale Up / Scale Down Testing: Measuring the Impact of Budget Changes

Scale Up and Scale Down tests let you measure the real, incremental impact of changing spend on a campaign that's already running. Instead of guessing whether more budget will drive more results, or whether you could pull back without losing much, you run a controlled test against a baseline and see the actual lift or loss.

This is the core building block for smarter budget allocation: once you know how efficiently a campaign converts each incremental dollar, you can shift spend toward the campaigns and subcategories where it does the most, and away from the ones where it doesn't.

Increase spend on a test campaign or geo above its current baseline, while a matched control group continues to run at the existing spend level. The difference in results tells you the incremental lift from the additional investment. In other words, it tells you whether the extra budget is actually working.

Decrease or pause spend on a test campaign or geo below its current baseline, while a matched control group continues running as-is. The difference in results tells you how much of your current performance that spend is actually driving. This is useful for identifying budget that isn't earning its keep.

Both test types compare against a baseline that's already running, so there's no need to launch anything new. You're measuring the marginal effect of a budget change on a campaign you're already investing in, with the campaign setup itself held constant. Only the weekly spend level differs between test cells.

Spend efficiency isn't the same across every campaign or subcategory. Some budget is highly incremental, while some is mostly redundant with demand that would have happened anyway. Scale Up and Scale Down tests give you a direct read on that difference, campaign by campaign.

Choose Scale Up/Down from the test type screen.

It's built to help you find your optimal budget level: the campaign setup stays the same, and only the weekly spend level differs between test cells. This is the option for budget optimization and scaling decisions, as opposed to Holdout, which is built for proving incremental ROI.

Pick the geography you want to test in.

Set your primary metric, which is Revenue by default.

Select your regional granularity.

Choose the campaigns and subcategories to include, filtered by tactic and channel. This is where you scope the test to the specific spend you're trying to evaluate rather than an entire account.

The wizard matches control geos to your test geos automatically, so the comparison reflects geos that would otherwise perform similarly.

Set the start and end dates for your test window.

Confirm the planned daily spend for each test cell alongside the selected geos and campaigns before launching.

Because the baseline is spend that's already running, results aren't a before-and-after comparison. They're a read on the marginal return of the spend change itself, measured against control geos that stayed at baseline.

For Scale Up, a positive incremental lift means the added spend is still generating results beyond what baseline spend was already driving, though marginal returns typically shrink as spend increases.

For Scale Down, incremental lift close to zero means the spend you removed wasn't adding much beyond baseline. This signals that the budget may be reallocatable.

Because Test Design lets you scope a test to specific campaigns and subcategories, you can run this at whatever level you're making budget decisions:

Use results side by side across those scopes to prioritize where the next incremental dollar should go and where existing spend could be trimmed without meaningfully hurting performance.

No. Scale Up and Scale Down tests run against a campaign that's already active. You're testing a change in spend relative to its current baseline, not starting from scratch.

Control geos are matched automatically using historical spend and revenue correlation, so they reflect geos that would be expected to perform similarly to your test geos without the spend change.

Each test needs its own test and control geo assignment, so run them as separate tests. Comparing results from both can help you find the efficient range for a campaign's budget.

Look for campaigns and subcategories with strong Scale Up lift as candidates for more investment. Look for campaigns with minimal Scale Down impact as candidates for reduced spend. Then, reallocate budget toward the areas where incremental dollars perform best.