How do I test branded search incrementality?

Usually by switching your brand ads off in part of your market and leaving them on in the rest. Then compare total sales and brand clicks, paid plus organic, between the two parts. If your free listing and Direct pick up most of what the ads used to claim, the campaign was collecting sales, not creating them.

Brand search is where ad reports flatter themselves most. Someone types your name, clicks the ad above your own listing and buys. The report books a sale. It never asks whether that buyer would have clicked the free result one line lower.

What one store's data shows

One store's anonymised GA4 export, 1 January 2024 to 21 August 2026. It holds shares of revenue only: no ad spend, no order counts.

First, where a buyer goes when the brand ad disappears. In that store's Channels sheet, Direct holds 57.7% of revenue in every view. By GA4's own definition, a Direct visit arrives through a saved link or a typed URL. A buyer who loses the ad can still walk in that way, or click your free listing. A campaign report sees neither, so it treats every paused click as a lost sale.

Then the clock. On the Journeys sheet, one-touch journeys carry 79.5% of revenue and take 0.5 days to buy. If your decided buyers look like that, a brand pause shows up in sales within days. That part of the test reads fast.

The slow lane does not. In that store's export, journeys of 2 to 3 touches take 12.5 days to buy, and 4 to 9 touches take 16.9 days. If a brand search is the last step of journeys like these, a one-week pause ends while some buyers are still deciding. Keep counting after the ads come back, for at least as long as your slow journeys take.

What the export cannot show is the brand campaign. Nothing in these rows isolates brand ads, and the file holds no spend. So it cannot say what brand ads cost that store or what they caused. It is one store, not a benchmark, and the answer for yours comes from a test on your own account.

Why does the usual answer mislead?

The usual answer is to pause the brand campaign for a week and see what happens. Right instinct, weak test. Four things go wrong.

• No comparison group. The paused week is set against the week before, with different weather, paydays and promotions. Google's researchers split regions at random into test and control, so both groups live through the same weeks. Without that, you are comparing against a guess.
• The wrong scoreboard. A paused campaign reports zero conversions by definition. What matters is whether total sales fall. Google's paid and organic report shows ad clicks, organic clicks and the two combined for each search, which is the scoreboard a brand test needs.
• Spend that moves instead of stopping. If Performance Max also serves on your name, pausing the brand campaign hands it those auctions. Google calls brand exclusions the most effective way to control how your ads appear on brand searches. A capped budget leaks too: Google warns that removing regions can push the savings into the control regions.
• A test that ends too early. Google's geo experiment guide wants pre-test data covering at least 3 times the test period. It also asks for a cooldown, with campaigns back on their usual budgets while you keep counting.

Research gives a range, not your answer. In 390 pause studies, Google's researchers found that, on average, 50% of ad clicks were incremental when the advertiser's own organic result ranked first. Field experiments at a large online marketplace found that brand-keyword ads had no measurable short-term benefits. Neither study is your store, which is the whole point of testing.

What can a brand test not tell you?

• What your brand demand is worth. People search your name because of everything else you do: products, email, word of mouth, other ads. Google even offers a Search Lift study, which measures whether campaigns lead to an increase in organic searches for your brand. A brand test values the ad on top of that demand, not the demand itself.
• Next quarter's auction. The result holds for the rivals bidding while you tested. Google's trademark policy does not restrict trademarks used as keywords, so a competitor or reseller can join later. When that happens, defending your name is worth more than your test said.
• Small effects. If brand ads add a sliver of sales, normal weekly noise can hide it. Google's geo lift tool shows a feasibility status before you start, and Google advises against proceeding when it reads Low. A flat result means "not seen", not "zero".

What to do this week

1. Write down your organic brand baseline. In Search Console, open Performance, click the filter bar, choose Query and select Branded. Pass: you can read weekly branded clicks going back months. Fail: the filter is missing, which Google says happens on sites with few impressions, so use Queries containing with your brand name.
2. Time the journeys your brand campaign closes. In GA4, click Advertising, open the Key events dropdown, choose Key event attribution paths and switch the table to Campaign. Pass: you can read Days to key event for paths that end in your brand campaign. Fail: brand clicks hide inside a mixed campaign, so split brand terms into their own campaign first.
3. See who else bids on your name. In Google Ads, open Campaigns, tick the brand campaign and select Auction insights. Pass: no rival or reseller shows up often beside you, so a pause test is cheap to run. Fail: one does, so test in fewer regions and expect your ads to defend real sales.

Check the homework. Your GA4 Attribution paths export already holds the evidence. Causality Engine reads that one file and shows what each channel caused next to what last-click gave it, in 1 to 2 minutes, for €99 once (excluding VAT), refundable within 30 days. Check the homework

Sources, 1 October 2026: Default channel group (Google Analytics Help); Measuring Ad Effectiveness Using Geo Experiments (Google Research); About measuring paid and organic search results (Google Ads Help); About brand exclusions (Google Ads Help); Implement campaigns for geo experiments (Google Ads Help); Impact Of Ranking Of Organic Search Results On The Incrementality Of Search Ads (Google Research); Consumer Heterogeneity and Paid Search Effectiveness: A Large Scale Field Experiment (NBER Working Paper 20171); About lift studies (Google Ads Help); Trademarks (Google Advertising Policies Help); Set up Conversion Lift based on geography (Google Ads Help); Performance report (Search results): Common tasks and use cases (Search Console Help); Performance report (Search results): Advanced filtering and comparison (Search Console Help); Key events attribution paths report (Google Analytics Help); Use auction insights to compare performance (Google Ads Help)

Related answers

• All Incrementality questions
• How to test branded search incrementality, step by step
• How do I test branded search for a food and drink store?
• Is branded search ad spend incremental? How to test it
• Incrementality testing, explained