Turning off search ads for their own brand names produced no statistically significant change in sales for eight small online retailers studied by marketing researchers. The result raises a question for businesses that judge search advertising mainly by the sales credited to a click.
The research, published in the Journal of Business Research, reports nine field experiments involving eight direct-to-consumer brands that sold exclusively online. Across the brands, researchers found no statistically significant sales effect from stopping branded search ads.
The findings are not proof that these advertisements never work. They show that, for the businesses and tests examined, the researchers could not detect a sales benefit. A different business, competitive setting or measurement period could produce another result.
Clicks and extra sales answer different questions
A branded search ad appears when someone searches for a business or product by name. A generic search ad targets a broader description, such as a type of mattress, without necessarily including a particular brand.
Someone who searches for a company by name may already intend to buy from it. If that person clicks a paid result and purchases, an advertising report can credit the sale to the ad. The buyer might nevertheless have reached the same site through an unpaid search result.
Marketers call the difference between what happened with an ad and what would have happened without it its incremental effect. That counterfactual cannot be observed for the same customer at the same instant, which is why researchers use comparison groups or controlled changes.
Our guide to click-through rate explains why the proportion of people who click is an early response measure, not proof that an advertisement caused an additional purchase.
What the researchers tested
Ilmira Beknazarova and co-authors examined small brands whose online sales were captured in the research dataset. The paper also reports a preregistered online experiment, in which the researchers specified the analysis in advance, looking at intentions to click.
For two of three brands that stopped a combination of generic and branded search advertising, website sessions fell significantly, but sales did not show a statistically significant effect. Visits and sales can move differently, so a change in traffic alone cannot establish what happened to revenue.
The study authors recommend considering generic search advertising or search engine optimization when allocating search budgets. Search engine optimization is work intended to make pages easier for search engines and users to find. It has costs of its own and is not guaranteed to replace paid visibility.
The available paper describes a relatively small set of online-only businesses. It does not establish the effect for local service providers, retailers with substantial offline sales, or a brand whose competitors bid heavily on its name. A non-significant result can also reflect uncertainty; it is not the same as proving that the true effect is exactly zero.
How a business can use the finding
Separate branded and generic campaigns in reporting, then define the decision before making a change. If a company tests spending reductions, comparable geographic areas or randomly assigned groups can offer a stronger comparison than simply judging one week against another.
Track total orders or qualified leads across channels, not just conversions attributed to the paused campaign. Allow enough time for delayed purchases, and consider seasonal demand, stock availability and competing promotions. A poorly powered test may be unable to distinguish a modest benefit from no benefit.
Cost still belongs in the calculation. A campaign that generates a small number of additional sales may be worthwhile if its incremental contribution exceeds its full cost. Conversely, a low reported cost per sale can conceal spending on customers who would have purchased anyway.
For businesses that depend on branded search, the result is a reason to test the budget with care. It is not a universal instruction to switch off an established campaign.
