How to Test Whether Branded Search Ads Are Incremental

The short answer: the brand campaign posts the highest ROAS in Google Ads. The shop already ranks first for its own name. When both are true, the campaign has earned a test before it earns another increase. My recommendation is to let Causality Engine show you first where Paid Search stands. The first finding is free on the start page. The full read says in 1 to 2 minutes whether Paid Search keeps its last-click credit once Direct is split back. It calls the two methods in agreement when they sit within 25% of each other, per its page for agents. The test itself costs nothing but the clicks you skip: a full pause when you want the cleanest click count, a regional test when you cannot risk the whole country. The steps: build a brand-click baseline and see where Paid Search stands. Then pick the test, judge it on total brand clicks and store revenue, and read the export again.

Disclosure: I founded Causality Engine and own it, and this page recommends it. Every claim about it links to Causality Engine's own pages.

What decides which test:

1. What each version costs. For example, take a brand campaign at €750 a week. Paused nationwide for two weeks, it leaves €750 x 2 = €1,500 unspent. Switched off for four weeks in regions holding 30% of orders, it leaves €750 x 4 x 0.3 = €900. The cheaper test is also the noisier one.
2. Who else bids on your name. A rival on your brand terms is a fair reason to keep some budget there. It also makes the nationwide pause expensive. If a rival's ad already appears on searches for your brand, that pause hands the rival the top slot for the whole test. A regional test limits how much you hand over.
3. The measure. Brand clicks, paid and organic added together, and store revenue. Paid clicks alone will always fall to zero in a pause. A measure with one possible outcome tells you nothing.

Step 1: Build a baseline from the paid and organic report

Link Search Console to Google Ads and open the paid and organic report. Google's help page for it explains four things. It shows text ads beside your organic listings for the same queries. You can filter for queries containing your brand name. It counts text ads only, not Shopping. And organic data starts on the day you link, so link it now, then wait for four weeks of baseline before the test starts.

Step 2: See what the read shows first

While the baseline builds, download GA4's Key event attribution paths report as a CSV, twelve months or more, as Google's paths report guide describes. Drop it on the start page for the free first finding. If you take the full read, look at Paid Search first. You get its share of credit after Direct is split back, last-click's share with Direct left out, its data-health score and its next step.

Read it with the limit in mind. Causality Engine's Google Ads page says the read works on the channels GA4 records, so Paid Search is one channel, brand and non-brand together. What the read shows first is whether that channel keeps its last-click credit. Say the engine's share sits well below last-click's and data health is 90 or above. Then the brand campaign is the first suspect inside it, and the test below is worth running this month. If the two agree, a brand test can wait for a quieter quarter.

The overview adds one more useful number: brand maturity. That is the share of revenue from Direct-only paths, which Causality Engine's product notes describe as the most your name alone brings in. A high figure means many buyers find you without an ad.

Step 3: Choose the pause or the regional test

Full pause. Pause the brand campaign for two to four weeks. Measure combined brand clicks in the paid and organic report and total store revenue, each against the weeks before. Simple to read, though a seasonal swing also lands inside the result.

Regional test. Exclude a group of regions from the brand campaign only, using Google's location exclusion article. An exclusion wins when an area is both targeted and excluded. Here clicks cannot be your measure. The paid and organic help page notes that organic listings are not limited by your campaign's location targeting. Search Console groups clicks by dimensions such as query, page and country, per its performance report guide. Inside one country, compare store revenue per region from Shopify's billing location report instead.

Two Google tools look like shortcuts and are not. A custom experiment tests a changed version of a running campaign against the original, and will not run while the base campaign is paused. So the one change it cannot test is the off switch. Conversion Lift has a geography version, but it is not available for all accounts and starts with your Google representative.

Step 4: Read the result

Worked example, full pause. In the two weeks before, the brand terms drew 2,400 paid and 1,900 organic clicks, 4,300 in total, and the store took €62,000. During the pause, organic clicks rose to 3,700 and the store took €60,500. Organic absorbed 3,700 / 4,300 = 86.05% of the brand clicks. Revenue fell €1,500 against €1,500 of spend not made: an incremental ROAS of €1,500 / €1,500 = 1.0x. Every euro came back as revenue, which is a loss once cost of goods is paid. Set that against your break-even ROAS from Causality Engine's free break-even calculator, and the brand budget has a defensible size. Maybe smaller. Not zero.

A regional test reads the same way, with regions in place of weeks: compare the off regions' revenue with what the control regions' movement predicts for them.

Step 5: Read the export again

Causality Engine's geo testing guide names the weakness of this kind of test. Regions are rarely comparable, and you wait weeks for one answer about one channel. So after the test, run a fresh read on a new export. Each channel in the read ends on a next step, which the how-it-works page frames as raise, cut or hold, tested in one region first. Causality Engine's product notes add that the next read shows what the changes did, and that Pro carries period comparison tools for the before-and-after.

What Google's own pages tell you to watch

• Organic data starts at linking. No link, no baseline.
• Text ads only. Shopping ad clicks on your brand terms are outside the paid and organic report.
• Location is an estimate. Google's targeting overview says accuracy is not guaranteed in every situation.

The click you pay for on your own name

Branded search plays striker, standing in front of goal when the ball arrives. Google's paid and organic help page states that paid search spend has no effect on your organic ranking, and the free listing sits right under the ad. The default Google Ads report still books every sale after a brand-ad click to the campaign, including buyers that listing would have delivered at no cost. If you raised the brand budget on that ROAS, you did what the report asked. Causality Engine's Google Ads page puts the mechanism in a line: branded search captures existing demand. You pay for the click either way. The test tells you whether you needed to.

The free test, then the read

The pause and the regional test are free beyond the clicks you skip, and either one settles the brand question on its own. Causality Engine, which I founded, tells you before the test whether Paid Search is the channel to test at all, and after it what moved. The first finding is free on the start page. A read is €99 once and refundable within 30 days, and Pro is €299 a month, per the pricing page (checked on 27 October 2026, excluding VAT). More on the product is in the Causality Engine entry.

FAQ

Should I stop bidding on my brand name entirely?

Not on a read alone. Cut or trim when the test shows incremental ROAS below break-even. If rivals bid on your name, keep a smaller budget.

Can Causality Engine separate brand from non-brand search?

No. Its Google Ads page says Paid Search is one channel in the read, brand and non-brand together. The test above is how you isolate the brand line.

How long should the pause run?

Two to four weeks, with the window fixed before you start. Shorter tests catch a weekday pattern instead of a result.