---
title: "Brand bidding: how we test whether your brand ads add sales"
url: https://marketingsignals.com/insights/brand-bidding-incrementality-test
type: Article
date: 2026-11-01
author: "Laura Salini"
---

# Brand bidding: how we test whether your brand ads add sales

Should you bid on your own brand? Our brand bidding test design: geo or time splits, what to measure, how long to run it and the rules for deciding after.

## Key takeaways

- Brand ads report strong returns because the people clicking were already looking for you, so platform figures cannot show whether the spend adds sales.
- Published research points both ways: eBay recovered 99.5% of paused brand clicks through organic results, while brands facing competitors lost 18% to 42% of clicks without ads.
- A geo split, with brand ads paused only in matched test regions, is the most reliable design; a time split is the fallback for smaller markets.
- Measure total brand clicks and total conversions in test and control regions, never ad clicks, and close the back door by excluding brand from Performance Max.
- Agree the decision rule before the test: keep brand ads where the cost per incremental conversion beats your target, and cut them where it does not.

Brand bidding is often one of the largest lines in a Google Ads account and one of the least questioned. The ads convert well, clicks are cheap and the reported return looks excellent, because the people clicking were searching for you by name.

The real question is how many of those customers would have clicked your organic listing anyway. Sometimes the answer is almost all of them. Sometimes a competitor is waiting in the ad slot above your listing, and the brand ad is protecting sales you would otherwise lose.

This article explains how we run a brand incrementality test, what to measure, how long to run it and the rules we use afterwards, so you can keep brand spend where it protects revenue and cut it where it buys traffic you already had.

## Why brand spend deserves a test

The published evidence points in two directions, which is exactly why a test is needed. When eBay switched off brand keyword ads on Yahoo and Microsoft's search engine, researchers found that 99.5% of the lost paid click traffic was captured by organic results instead ([Blake, Nosko and Tadelis, NBER](https://www.nber.org/papers/w20171)). A follow-up test on Google saw total Google-referred traffic fall by 3%.

A later study of thousands of brands on Bing found a different picture where competitors were present. With no rivals in the auction, a brand's own ad added only 1% to 4% more clicks. Brands that did not advertise while competitors bid on their name lost 18% to 42% of their clicks to those competitors ([Simonov, Nosko and Rao, Marketing Science](https://ideas.repec.org/a/inm/ormksc/v37y2018i2p200-215.html)).

Competitors are allowed to do this. Google's [trademark policy](https://support.google.com/adspolicy/answer/6118) does not restrict the use of trademarks as keywords. Restrictions apply to how a trademark is used in the ad itself, and Google reviews those when the trademark owner complains.

For the business, not testing is costly in both directions. Cut brand spend blindly and a competitor can take customers who were already looking for you. Keep it blindly and you may pay for clicks your organic listing would have won for free. Brand campaigns also flatter a blended return on ad spend, which can make non-brand campaigns look healthier than they are, so our [pay per click](https://marketingsignals.com/service/pay-per-click) reporting always separates the two.

## How we run a brand incrementality test

A brand test is a form of [incrementality testing](https://marketingsignals.com/glossary/incrementality-testing): it compares what happens with the ads against what happens without them. The design matters more than the tools.

Google Ads offers Conversion Lift based on geography, which supports Search campaigns and splits a single country into regions that see ads and regions that do not ([Google Ads Help](https://support.google.com/google-ads/answer/14097193)). It is not available on every account and is arranged through a Google account representative, so we use it where it is offered and run the design below where it is not.

1. Split brand terms by what the searcher wants. Pure brand searches, brand plus product, and brand plus navigation (login, contact, returns) behave differently, so we test them as separate groups. Someone searching for your login page rarely needs an ad. Brand plus a product name may face resellers and competitors.
2. Check who is in the auction before you start. We pull auction insights for the brand campaign to see which advertisers appear and how often their ads overlap with yours, and search the terms on mobile and desktop in the markets that matter. Heavy competitor presence tells us to expect a loss and measure it carefully, not to skip the test.
3. Close the back doors. If you pause the brand campaign, Performance Max and broad match keywords in other Search campaigns can start serving on the same searches, and the test then measures nothing. We apply brand exclusions to Performance Max and non-brand Search campaigns ([Google Ads Help](https://support.google.com/google-ads/answer/13721847)) and check the search terms report daily in the first week.
4. Use a geo split where you can and a time split where you cannot. In a geo split we divide the country into test and control regions with similar historical brand searches and conversions, then pause brand ads only in the test regions. Google Ads can target regions, cities or postcodes depending on the country. Seasonality and news hit both groups alike, so the gap between them is the effect of the ads. Where the market is too small to split, we run brand ads on, then off, then on again for equal periods, and compare with the same weeks last year.
5. Measure total brand traffic and conversions, not ad clicks. Ad clicks fall to zero in the test regions by design, so they say nothing. What matters is total brand clicks (paid plus organic) and total conversions and revenue from those regions. Search Console reports by country, not by region, so for a geo split we use city data in GA4 or postcodes from the CRM. In the UK, GA4's region dimension is broad, so city is usually the better level for a split.
6. Run it long enough to cover the buying cycle. We plan for at least four weeks with brand ads off, after a pre-period of the same length that confirms the test and control groups move together. We avoid peak trading weeks unless peak behaviour is the question, because a short or unusual period easily reads normal weekly noise as a result.
7. Agree the decision rule before the test starts. We calculate the cost per incremental conversion: brand spend saved in the test regions divided by the conversions lost against control. If it is below the target cost per acquisition for new customers, the brand ads are earning their place. If it is well above, we cut or reduce them. Writing the threshold down first stops anyone reading the result to suit their budget.
8. Act term by term, then test again. The outcome is rarely all or nothing. Typical results are keeping pure brand where competitors appear, cutting navigational terms, or running brand ads only during promotions when the ad carries the offer. We repeat the test every six to twelve months, or sooner when a new competitor shows up in auction insights.

## How to measure the result and what to report

The report upwards should fit in one sentence: brand ads cost this much a month and bring in this many conversions that would not otherwise have happened, at this cost each, against our target. That turns brand spend from an assumption into a decision the board can see. Behind that sentence sit five measures.

- Total brand clicks, paid plus organic, in test and control. If organic picks up nearly all the lost paid clicks, the ads were mostly duplicating free traffic.
- Total conversions and revenue in test and control, from the CRM or back end where possible, since each platform attributes differently.
- Cost per incremental conversion, set against the target cost per acquisition for new customers.
- Competitor presence on brand terms during the test, from auction insights, to explain any loss.
- New versus returning customers lost, where the data allows. A brand ad that only reaches existing customers is worth less than one that wins new ones.

## When a test does not fit, and the mistakes we see most

Brand testing is a standard part of our [integrated search](https://marketingsignals.com/service/integrated-search) work, because the right answer depends on the [paid and organic overlap](https://marketingsignals.com/glossary/paid-and-organic-overlap) for each term. It does not suit every account, and the same errors come up often.

- Low brand search volume. If brand campaigns bring a handful of conversions a week, four weeks cannot separate a real effect from noise. Run a longer time split or accept a less certain answer.
- Your organic listing is not first for your own name. If resellers, marketplaces or review sites outrank you, the brand ad is doing a different job, and the fix starts in organic.
- Measuring ad clicks or platform-reported conversions as the outcome. Both disappear by design when the ads are paused.
- Leaving Performance Max free to serve on brand searches during the test, so it quietly replaces the paused campaign.
- Testing all brand terms together, so a loss on contested brand plus product terms hides a saving on navigational ones.
- Asking whoever is paid on media spend to design and judge the test alone. A positive result can cut their fee, which is a reason to agree the method and decision rule in writing first.

## Questions to ask your team or agency

- What share of our paid search spend and reported conversions comes from brand terms?
- When did we last test whether brand ads add conversions, and what did we measure?
- Which competitors appear on our brand terms, and how often?
- Are Performance Max campaigns excluded from brand searches, and if not, how are those conversions reported?
- What cost per incremental conversion would make us keep brand spend?

## Frequently asked questions

### Should I bid on my own brand name in Google Ads?

It depends on who else is bidding. Research on eBay found organic results recovered 99.5% of the clicks lost when brand ads were paused, while a Bing study of thousands of brands found that brands not advertising lost 18% to 42% of clicks when competitors bid on their name. The reliable answer for your brand comes from a geo or time-split test that measures total conversions.

### Can competitors bid on my brand name?

Usually, yes. Google's trademark policy does not restrict trademarks used as keywords, so a competitor can target searches for your brand. Use of your trademark in the ad text itself is different: if the trademark owner submits a complaint, Google reviews it and may restrict that use. Auction insights in Google Ads show which advertisers appear on your brand terms and how often.

### How long should a brand bidding test run?

We plan for at least four weeks with brand ads paused, after a pre-period of the same length that confirms test and control regions behave alike. Run longer for brands with low search volume or long buying cycles. Avoid peak trading periods unless that is what you want to measure, and do not judge the result in the first week, when normal variation looks like an effect.

## Sources

- [Blake, Nosko and Tadelis: Consumer heterogeneity and paid search effectiveness, NBER working paper 20171 (published in Econometrica, 2015)](https://www.nber.org/papers/w20171)
- [Simonov, Nosko and Rao: Competition and crowd-out for brand keywords in sponsored search, Marketing Science (2018)](https://ideas.repec.org/a/inm/ormksc/v37y2018i2p200-215.html)
- [Google Research: Studies show search ads drive 89% incremental traffic (2011)](https://research.google/blog/studies-show-search-ads-drive-89-incremental-traffic/)
- [Google Research: Impact of organic ranking on ad click incrementality (2012)](https://research.google/blog/impact-of-organic-ranking-on-ad-click-incrementality/)
- [Google Ads policy: Trademarks](https://support.google.com/adspolicy/answer/6118)
- [Google Ads Help: Use auction insights to compare performance](https://support.google.com/google-ads/answer/2579754)
- [Google Ads Help: About brand settings for Search and Performance Max](https://support.google.com/google-ads/answer/13721847)
- [Google Ads Help: Target ads to geographic locations](https://support.google.com/google-ads/answer/1722043)
- [Google Ads Help: Set up Conversion Lift based on geography](https://support.google.com/google-ads/answer/14097193)
- [Google Analytics Help: Analytics dimensions and metrics (Region)](https://support.google.com/analytics/answer/9143382)
- [Search Console Help: Performance report dimensions and data groupings](https://support.google.com/webmasters/answer/17011259)

Visibility Playbooks

## Related Playbooks

More Insights

## More articles we think you’ll like