Sixty-five percent. That's how much of the paid-attributed revenue came back through organic and direct after one company turned off $113,000 a month in paid search. The test ran thirteen weeks. Branded search never came back on.
Anna Crowe's experiment at Search Engine Land is the kind of controlled spend-down that most marketing leaders talk about running but never actually execute. The fear is rational: what if revenue drives off a cliff while the CFO watches? So paid stays on. Forever. The budget becomes a recurring line item nobody questions because questioning it means owning the downside.
But the math from this test deserves a seat in your next pipeline review.
The Incrementality Question Nobody Wants to Model
The company paused branded paid search across the US, UK, Australia, and Canada. By month's end, most non-brand paid was off too. The monthly spend had been running between $106,000 and $119,000 across branded search, non-brand search, Shopping, and Performance Max.
Here's the number that should make you pull your own data: nearly nine out of ten dollars in the branded campaign were defending clicks that organic was already positioned to win. The business wasn't acquiring customers. It was paying Google for traffic it would have captured anyway.
This is the incrementality problem that keeps attribution vendors employed and CMOs nervous. Google's own research, cited in Crowe's analysis, studied more than 400 advertisers who paused search campaigns and found that 89% of ad clicks were incremental on average. But "on average" is doing a lot of work in that sentence. Your branded terms, where you already rank first organically, are not the same as competitive non-brand queries where you're fighting for position six.
The distinction matters because it's the difference between buying customers and buying your own clicks.
What the 13-Week Window Actually Showed
The test wasn't a clean binary. Crowe documents the "what ifs" that make these experiments hard to run: What if the ad at the top builds trust? What if paid plus organic together creates more SERP real estate? What if competitors start bidding on your brand name? What if turning off paid doesn't move the click to organic but makes it disappear entirely?
These are real concerns. VI Marketing's analysis of paid and organic synergy makes the case that appearing in both positions creates compounding visibility, not redundant spend. When your brand shows up twice on the results page, you're harder to miss.
But the test results suggest that for this company, the synergy argument was worth less than $113,000 a month. The 65% revenue recapture through organic and direct channels means the actual incremental value of paid was closer to 35% of what the channel report claimed. That's a $73,000 monthly gap between attributed value and real value.
Run that number through your own CAC payback model. If a third of your paid search budget is buying clicks you'd get for free, your effective CAC is higher than your dashboard shows, and your payback period is longer than you're telling the board.
The Compounding Asset Argument
Fractl's research on organic versus paid ROI found that 99% of consumers used a search engine in the past week, while only 5% say they click on paid ads "always" or "often." The consumer preference data is stark, but the more interesting finding is structural: organic rankings keep driving traffic after publication, while paid visibility stops the moment the budget does.
This is the compounding asset argument that SEO teams have been making for years. PUSH Group's 2026 analysis puts it bluntly: organic search drives 53% of all website traffic, paid drives 27%, and when you stop paying, one of those disappears.
The counterargument is speed. Paid gives you visibility today. Organic takes months to build. For product launches, seasonal campaigns, or competitive conquesting, paid search is the right tool. The mistake is treating it as the whole strategy instead of a bridge while organic catches up.
Running Your Own Test Without Driving Off the Cliff
Most companies won't pause $113,000 in monthly spend to find out what's incremental. The career risk is asymmetric: if you're wrong, everyone knows; if you're right, you saved money that was already in the forecast.
But you can design a smaller experiment. Pick one market, one product line, or one branded term cluster. Run a four-week holdout. Measure total revenue from search (paid plus organic plus direct) before and after, not just paid-attributed revenue. The goal is to see whether the pie shrinks or whether the slices just move.
If you're running Performance Max or broad match campaigns, the incrementality question is even harder to answer because Google's automation optimizes for attributed conversions, not incremental ones. The algorithm will happily spend your budget on clicks you would have won organically if that's what drives the conversion number up.
The Board-Ready Version
Here's how I'd frame this for a pipeline review:
We're spending X on paid search. Channel attribution says it's driving Y in revenue. But incrementality testing suggests Z percent of that revenue would come through organic anyway. Our real CAC on paid search is not X divided by conversions; it's X divided by incremental conversions. If we reallocate the non-incremental spend to channels with higher incrementality, we improve CAC payback by W weeks.
The Crowe test gives you a benchmark: 65% recapture through organic, 35% true incrementality. Your numbers will differ based on brand strength, organic rankings, and competitive dynamics. But if you're not running the test, you're assuming 100% incrementality, and that assumption is almost certainly wrong.
The $113,000 question isn't whether paid search works. It's whether you're buying customers or buying your own clicks. The only way to know is to turn something off and watch what happens to the total, not just the channel.
