Should you bid on your own brand? What the studies foundShould you bid on your own brand? What the studies found

Yourbrand spends about £3,600 a month on clicks from people who searched for “yourbrand”. The finance director asks a fair question: they were already looking for us, and we're first in the organic results, so wouldn't they have clicked that link instead?

There's published research on exactly this, and it's often quoted as a one-line answer. It isn't one. The studies disagree, and the reasons they disagree are the useful part, because they tell you which conditions matter. This guide goes through what each found, what changes the answer, and how to test your own brand properly.

What the question really is

The question isn't whether brand ads get clicks. They do, lots of them, cheaply. It's how many of those clicks, and the orders behind them, wouldn't have happened without the ad. That share is the ad's incrementality. A click that would otherwise have gone through your organic listing costs you money and adds nothing; a click that would otherwise have gone to someone else's ad is the one you're paying for.

So every study below is measuring substitution: when the ad goes, how much of its traffic comes back some other way, mostly through the organic result.

The studies, one by one

Blake, Nosko and Tadelis (2015). In March 2012 a very large US online marketplace stopped bidding on every search containing its name on Yahoo! and Bing, and kept bidding on Google, which the authors used as the control. Measured against Google, the clicks reaching its site from those engines barely moved: the authors estimate that 99.5% of the traffic the brand ads had brought was retained, nearly all of it through the organic link. A later test in Germany switched brand ads off in 8 of the 16 states, chosen at random, and found no measurable effect on sales.

The authors are clear about the limits. The marketplace was one of the best-known names online, few rivals were seen bidding on its name once it stopped, and the effects measured were short-term. They say the result may not hold for smaller or newer firms. It's the study most often quoted, and the one least like most brands.

Chan, Yuan, Koehler and Kumar (2011). Google's researchers looked at 446 cases where advertisers had themselves paused or sharply cut their search ads, in the US, Germany, France and the UK (34 of them British), and modelled how many paid clicks organic clicks made up. On average about 89% of paid clicks were incremental. That covers every kind of keyword, not just brand terms, the pauses weren't randomised, and the advertisers chose when to pause. The authors recommend randomised tests.

Chan, Kumar, Ma and Koehler (2012). Google's follow-up split 390 pause studies by where the advertiser's own site ranked organically for the search. With no organic listing on the first page, 99% of ad clicks were incremental; with the site first organically, about 50%; at positions 2 to 4, about 82%. Brand searches are usually the ones where you rank first, so this is the figure closest to the brand question. It measures clicks, not sales.

Simonov, Nosko and Rao (2018). A randomised experiment on Bing in January 2014, changing how many ads showed above the results on searches where the brand ranked first organically, across 2,517 brands. Where no rival advertised, the brand's ad added on average about 2.3 clicks per 100 searches, and less for more prominent brands. Where the brand didn't advertise and a rival's ad sat on top, the brands lost 12% of their clicks on average, and for the highest-traffic brands a single rival took about 18%.

Coviello, Gneezy and Goette (2017). A US car research website switched off its brand ads on Bing and Yahoo in half of its US regions for about three weeks in late 2015, at a time when rivals were bidding on its name. It lost roughly half to 60% of the traffic the brand ads had brought, and most where it had relied on paid clicks most. The measure was website sessions, not sales, and it's a working paper.

Golden and Horton (2021). A useful counterweight. Two online services marketplaces of similar size; one switched off all its Google search ads in half the US for a month. When its ad left its own brand search, the rival's ad moved up about one place, but the rival gained almost no extra clicks or sign-ups from it.

The numbers are made up, but the sum is the one Simonov and colleagues used. Ten times the price per click is the same spend seen honestly. Whether £1.80 is cheap depends on what an extra visitor is worth to you, and on what happens to the other 18,000 when someone else's ad sits above your organic link.

What changes the answer

• Who else bids on your name. Every study with rivals present found more at stake than the ones without. On your brand that includes PPC affiliates your terms allow, ones they don't, voucher pages and resellers.
• Where you rank organically. First place, with sitelinks, takes back most of the clicks; further down, much less comes back.
• How well known you are. The effects were smallest for the most prominent brands in every study that measured it.
• Which searches. “yourbrand” and “yourbrand discount code” are different auctions with different advertisers. The studies mostly looked at the bare name.
• The page itself. Since 14 April 2025 Google runs the top and bottom ad locations as separate auctions, with one ad per business or site in each. An ad that leads to your shop, such as an affiliate's direct linking ad, competes with yours for the top.
• What the click costs you elsewhere. If an affiliate's ad takes the click, the order can carry commission under last click rules, which none of these studies counted.

How to run your own test

The method the better studies share is a geo experiment: switch brand ads off in some regions and leave them on in the others, then compare all your orders, not just paid clicks. Google's own researchers describe the design in Vaver and Koehler's 2011 paper on geo experiments.

1. Write down the measure before you start: orders and revenue from every channel in each region, with organic clicks from Search Console and affiliate commission beside them. Paid clicks alone will always fall.
2. Record who appears on your brand searches now, on the bare name and the code searches, by day and by night, and read your brand campaign's Auction insights. This is your baseline for who might move in.
3. Split the UK into regions you can target and that your sales data can see, such as groups of postcode areas. Match them into pairs on past sales and pick one of each pair at random to go dark. Ranking by size before randomising narrows the result.
4. Run a pre-period of several weeks with nothing changed, to check the two groups move together. Then switch off for at least two weeks, longer if your customers take time to buy. Avoid a sale, a launch or Christmas.
5. Exclude the test regions from your brand campaigns with location targeting. Exclusions work on where Google thinks the searcher is, so expect some spill at the edges.
6. Keep your programme terms and affiliates' permissions unchanged, and keep recording who appears on your name in the test regions. A new advertiser there is part of the result.
7. Compare total orders in the test regions with the control regions, against how they compared in the pre-period. Then look at what moved: organic clicks, direct visits, and commission on sales that began with a brand search.
8. Decide, and plan to test again. Rivals and affiliates change, so a result from two years ago is a guess about today.

Two Google Ads reports help. The paid and organic report, once Search Console is linked to Google Ads, shows by search query how often your ad, your organic listing or both appeared and were clicked. It shows the overlap, not the cause, but it tells you which brand searches you rank for organically before you touch them. Google Ads also offers lift studies to some accounts; whether one is open to yours is a question for your Google contact.

The ordinary explanations

A test throws up things that look alarming and usually aren't. Most of these are the test working as designed.

• Organic clicks rose in the test regions. That's the substitution the test is measuring, not a fault.
• Orders fell in one region for a reason of its own: weather, a local promotion, a delivery problem. That's why you compare against a matched region and a pre-period.
• A new advertiser appeared on your name in the test regions. Someone may simply have bid on an auction you'd left; an approved partner, a reseller or your own agency are the first to check.
• Search Console's figures arrive a couple of days late, and it splits data by country, not region, so organic clicks by region need your own analytics.
• Your ad still showed in a test region. Location estimates are mostly by IP address, and a phone on mobile data can be placed by its carrier's address.

Before you switch off for good

Whatever the test says, write down what your programme terms allow on your brand searches if your own ad isn't there. Many terms were written assuming the brand always bids, and a partner who reads “don't outbid the brand” may reasonably take an empty top spot as fair. Then keep watching your name, nights and weekends included, because the people who move into a gap don't always do it during office hours. The guide to what others' ads on your brand cost puts a figure on what a gap can cost in commission, using these studies' numbers.