Someone Paused $113,000 a Month in Paid Search. Before You Do the Same at Your Store, Read What the 89% Actually Means.

Yesterday Search Engine Land published something most agencies would never let a client read.

An SEO consultant named Anna Crowe wrote up what happened when a company turned off $113,000 a month in paid search. Not trimmed. Off. Branded campaigns paused across the United States, United Kingdom, Australia and Canada, and by the end of that month most of the non-brand paid search was off too. Branded search never came back on.

Then she did the part almost nobody does. She matched every Google Ads search term and its spend against every matching Search Console query and its organic clicks, keyword by keyword, over the same period, and sorted each query into three buckets: clicks organic was already positioned to win, clicks where organic performance actually depended on the ads, and clicks paid genuinely reached that organic could not.

On one branded campaign carrying $36,129 in spend, $3,945 bought something organic could not have captured. That is 10.9%. The other $32,184 was brand defense. On queries where paid and organic both showed up, organic was already taking 71% of the clicks with the ads still running.

Thirteen weeks after the pause, organic and direct had recaptured 65% of the revenue the paid channel used to get credit for.

If you run a franchised store, you have a branded campaign. Somebody set it up years ago, nobody has questioned it since, and it quietly bills every month. This article is going to land in your inbox from a vendor or a 20 Group buddy with a subject line telling you to shut yours off.

Do not shut yours off. Read the next part first, because the reason this test is interesting for dealers is also the reason its answer does not transfer to your store cleanly.

The evidence has been split for fifteen years

The uncomfortable thing about brand defense is that the two most famous experiments ever run on it reached opposite conclusions, and both were credible.

In 2011, four Google statisticians published Incremental Clicks Impact of Search Advertising in the Journal of Advertising Research. They looked at more than 400 advertisers who had paused search campaigns and measured what happened to organic clicks in the gap. Their finding: more than 89% of ad clicks were incremental. Organic did not come pick up the slack. Turn the ads off and most of those visits simply did not happen.

Then in 2014, three economists ran the other direction. Thomas Blake, Chris Nosko and Steven Tadelis had eBay turn off paid search at scale and published the results as Consumer Heterogeneity and Paid Search Effectiveness, later in Econometrica. Brand keyword ads showed no measurable short-term benefit. For a name everyone already knew, the ads were charging admission to customers walking through the door on their own.

Notice the number that shows up on both sides of this fight. Google measured 89% incremental. Crowe measured 89.1% defense. Same figure, opposite meaning, and the difference between them is not methodology. It comes down to who was running the ads.

eBay had overwhelming organic ownership of its own name. The 400-plus advertisers in Google's dataset, on average, did not. Brand defense is worth exactly what your organic position fails to capture on its own, which is a different number for every business on earth.

Including yours.

Why your dealership is the harder case

Here is where the automotive version diverges, and it cuts both ways.

Your brand auction is contested in a way eBay's was not. Google's own trademark policy lists using trademarks as keywords among the things Google Ads will not restrict. The restrictions apply to trademark use in the ad text, and even then the trademark has to appear in the ad rather than only on the landing page. Translated: anyone can bid on your dealership's name tomorrow morning and Google will not stop them. Third-party marketplaces, the store across town, an OEM tier-2 campaign, and your own used-car aggregator feed can all end up in your brand auction. When eBay went dark on its own name, the results page mostly went quiet. When you go dark on yours, the space above your organic listing does not stay empty.

Do not take my word for who is in there. Open Auction Insights in your Google Ads account, filter to your branded campaign, and look at who has impression share against your own name. That report will tell you in five minutes whether your defense line has anyone on the other side of it.

Your organic brand position is usually strong, which cuts the other way. A franchised dealership that has been in a market for twenty years and has a healthy Google Business Profile typically owns the top of its own name comprehensively: the map pack, the site link block, the whole first screen. That is the exact profile where Crowe's finding is most likely to hold, and where every dollar of brand spend is most likely to be buying a click you already had.

Your conversion rates are not equal, and the gap matters more here. In her test, paid traffic converted at about 2.9% and organic at about 2.3%. That gap is real, and at a dealership it is often wider, because a branded paid click can land on a hand-built offer page while the organic click lands on a homepage. If you pause paid without fixing where organic lands, you inherit the worse landing experience along with the cheaper traffic.

Some of your brand spend is not your money. A meaningful share of dealer branded search runs through co-op, or through an OEM-mandated tier-2 program you do not control the switch on. Pausing spend that a manufacturer reimburses does not improve your P&L. It forfeits a subsidy. We wrote about how co-op certification is tightening, and the same logic applies here: know whose dollars are in the campaign before you touch it.

And the ground under the organic half is moving. Crowe's whole test rests on an assumption that has been safe for twenty years: pause the ad and the click drops down to your organic listing. That assumption is getting weaker every quarter as AI Overviews and AI Mode absorb clicks that used to reach a blue link. We covered research suggesting AI Overviews take a disproportionate share of high-intent traffic, and it is the biggest reason a 2014 eBay result should not be pasted onto a 2026 dealership. The recapture rate you would get today is an open question, and it is measurable only in your own account.

What to actually do about it

The wrong takeaway is to cancel branded search on Monday. The right one is that "we have always run branded search" stops being an answer once someone asks what it costs. Crowe's line on that is blunt and worth repeating: "We've always run branded search" isn't a strategy.

Here is the sequence that works at a store.

1. Link Search Console to Google Ads and open the paid and organic report. This is Google's own tool and most dealer accounts have never turned it on. It splits every query into three states: organic listing only, ad only, and both showing together. The "both" rows are your defense line, sitting there in your own account. You do not need a consultant to run this.

2. Price the defense. Take last month's branded campaign spend and figure out what share of it landed on queries where your organic listing was already on the page. That single percentage is the number this whole debate is about, and almost no dealership knows theirs.

3. Check Auction Insights before you conclude anything. If competitors and marketplaces are showing on your name at meaningful impression share, your defense spend is doing real work. If the auction is empty, it is a subscription.

4. Separate co-op and OEM tier-2 money from your own. Only your money counts as savings.

5. If you test, test geographically and give it a quarter. Pause branded in one or two comparable markets, keep the rest running, and hold the line for at least twelve weeks. Crowe's recapture went from 30% at six weeks to 65% at thirteen. A four-week test would have declared a disaster and turned everything back on.

6. Fix your organic landing experience before you shift traffic to it. If your brand-term organic result drops shoppers on a generic homepage while your ad drops them on a live offer, you have a landing problem wearing a channel problem's clothes.

7. Judge it on gross, not on attributed revenue. This is the part of Crowe's writeup that deserves to be printed and taped to a wall. Top-line revenue fell roughly $30,000 to $40,000 a month after the pause, and the working estimate still put net P&L $5,000 to $20,000 a month higher, because the media spend went away with it. Her summary of why the dashboards missed that: "Channel dashboards are good at telling you what revenue an ad touched. They're less good at telling you whether you needed to buy the touch in the first place."

The honest caveats

Two, and they matter.

This is one anonymized case study from one unnamed company in an undisclosed industry, published on a site owned by Semrush and written by a contributor. It carries nothing like the weight of Google's 2011 work across 400-plus advertisers. The methodology is sound and shown openly, and one company is still one company.

The test also was not automotive. The query mix in it was product terms and category plurals. A dealership's brand auction has franchise names, service queries, and third-party inventory sites tangled into it in ways a direct-to-consumer brand never deals with.

What survives both caveats is the method, and the method is free. Match your paid spend to your organic clicks, query by query, and find out what share of your branded budget is buying something you did not already have.

Most dealership marketing arguments are about which vendor to fire. This one is about a line item that has never been asked to justify itself. We have written before about why dealer paid search keeps getting more expensive. The cheapest response to rising costs is usually to stop paying for the traffic you were going to get anyway.

If you want us to run the paid and organic split on your store's account and tell you what your defense number actually is, book a walkthrough and we will go through it with you.