Twenty-one thousand. That’s the daily organic visits one SEO drove to a test page by manipulating clicks. The most rigorous public attempt on record. Rankings didn’t move. Two months later, they dropped.

That test matters right now for a specific reason. Google confirmed in a courtroom and in a leaked API that clicks genuinely do feed its ranking system. Every click-farm vendor selling “CTR boosting” in 2026 opens with that same true fact. Then they close on a conclusion the evidence doesn’t support. That renting the signal works.

In October 2023, Google VP of Search Pandu Nayak testified in the DOJ’s antitrust trial about a system called NavBoost. He described it as a system that memorizes past clicks, trained on user data, running on a rolling window of roughly 13 months. That testimony ended a decade of SEO forum debate that any law firm SEO program has had to work around ever since. NavBoost isn’t a rumor. It’s sworn testimony in a federal courtroom.

Seven months later, the picture got sharper. The May 2024 Content Warehouse API leak surfaced the actual field names inside Google’s ranking infrastructure. goodClicks. badClicks. lastLongestClicks. unsquashedClicks. The last one is the tell. If there’s an “unsquashed” version of a click count, something else squashes it.

The Leak Named the Function Built to Catch You.

That squashing function normalizes and down-weights abnormal click patterns before they influence anything. Google hasn’t published exactly how it works. It doesn’t need to. The field’s existence alone tells you something. Google already assumes some of its click data is manufactured, and built a step to discount it, one more reason the fundamentals in a real technical SEO checklist still matter more than gaming a single signal.

Leaked Content Warehouse API · concept diagram

The leak named the fields Google built to catch a rented click.
goodClicks badClicks lastLongestClicks unsquashedClicks squashing function ranking

Concept relationship, not measured data · May 2024 Content Warehouse API leak, via SparkToro

Read that again next to any pitch that says “the leak proves clicks are a ranking factor, so buy some.” The same leak that proves clicks matter also proves Google built a filter for clicks that don’t look real. The premise is true. The vendor’s conclusion is the part that fails.

One Honest Test, Two Months of Nothing, Then a Drop.

Most CTR-manipulation claims run on private case studies with no methodology and no losses to report. Bartosz Góralewicz at Onely ran the opposite. A public, documented experiment on a live page, ugly result published right alongside the good one.

He manipulated search behavior on a target keyword until Google’s own Keyword Planner registered the volume as real. Organic traffic to the page peaked at roughly 21,000 visits a day. If clicks moved rankings the way vendors claim, that page should have climbed. It sat still for two months. Then rankings declined.

The Onely CTR test · illustrative timeline

Clicks hit ~21,000 a day. Rankings sat still for two months. Then they fell.
Month 0 Month 1 Month 2 Month 3 ~21,000 visits/day (illustrative peak) baseline Rankings: no movement Rankings: declined Organic clicks (rented) Google ranking status

Illustrative depiction of the reported outcome, not Onely's raw daily data · Onely, via Search Engine Land

Around the same period, Google’s Gary Illyes addressed the mechanic directly at a search marketing conference. He said Google uses SERP clicks “for evaluation and for experimentation, but not for ranking.” One sentence, from the company running the system, matching one public test that actually tried to prove the opposite and failed.

Detection Doesn’t Look Like a Penalty. It Looks Like a Refund.

Nobody selling click farms is naming a client Google slapped with a manual action for it. That’s not because it doesn’t get caught. It’s because the typical outcome isn’t a penalty screen. It’s a quiet reversion.

What Buying Clicks Actually Buys You
~13 months NavBoost's rolling click-history window, per DOJ trial testimony Source: Pandu Nayak, USA v. Google, Oct. 2023
0 lift, 2 months Ranking change during Onely's 21,000-visit-a-day CTR test Source: Search Engine Land
Discount, not a flag Manipulated signals typically get down-weighted, not manually penalized Source: Leaked Content Warehouse API fields

A rank that climbed on rented clicks tends to fall back the moment the spend stops. The underlying page never earned the behavior the clicks were faking, which is why measuring legal advertising effectiveness has to track signed cases, not a click count.

What’s left afterward isn’t neutral. It’s wasted media spend, a traffic spike analytics can’t explain, and sometimes an ad account carrying a fraud flag it didn’t have before.

Some vendors in this market also sell the mirror version. “Negative CTR” campaigns that click on a competitor’s listing to drag it down. Same mechanism, same detection risk, aimed the other way.

AI Overviews Shrank the Prize You’re Renting.

Even in a world where buying clicks worked, the payoff would be worth less than it was two years ago. Ahrefs tracked position-1 organic click-through rate across 300,000 keywords. On AI-Overview-triggering terms, CTR fell roughly 58% by December 2025, up from a 34.5% drop the same team measured in April 2025, part of the same shift our own AI legal-citation study tracked on the citation side.

Position-1 organic CTR · AI-Overview keywords

AI Overviews cut the click prize further every time Ahrefs checks.
0% 20% 40% 60% −34.5% April 2025 −58% December 2025 23.5 more points, worse

Ahrefs, 300,000-keyword study · position-1 organic CTR on AI-Overview-triggering keywords

Position one used to be the reward. Now a growing share of that traffic never leaves the results page at all, the same trend covered in AI Overview citations lifting paid clicks. Buying artificial clicks to chase a ranking spot pays out against a shrinking pool of real ones. And it pays out on a signal the system already distrusts.

For Regulated Advertisers, the Downside Compounds.

Every advertiser buying fake engagement risks wasted spend and a fraud flag. Licensed professionals risk their license. The FTC’s Operation AI Comply, launched in September 2024, already prosecutes fake reviews and manufactured engagement as deceptive practices. That’s a business-wide exposure, not a legal-industry one.

For an attorney, it stacks on top of a professional conduct code that shapes how legal marketing gets built in the first place. ABA Model Rule 7.1 prohibits false or misleading communications about legal services. Rule 8.4(c) makes dishonest conduct professional misconduct, full stop.

A firm that pays for fake clicks to inflate its visibility isn’t running a gray-area growth tactic. It’s building a paper trail that reads, in plain language, as the exact conduct both rules exist to catch.

Law is the sharpest version of this risk. It isn’t the only version. Any brand buying a fake signal is one audit away from explaining it to someone who controls its distribution.

The Signal Is Real. Renting It Is the Losing Move.

Here’s the shape of it. Google confirmed clicks matter. That confirmation became the opening line of every pitch selling you a shortcut to more of them.

The only public test rigorous enough to check the pitch found nothing to buy. Zero lift for two months, then a loss. A system built with fields named specifically to catch abnormal click patterns isn’t waiting to be fooled. It’s already built for this exact fraud.

This is the same pattern we flagged in Reddit becoming the new search results. A real, measured signal, and a market rushing to fake the thing that made it valuable. Faking a real signal doesn’t beat the system that measures it. It just tells that system exactly where to look.

If you want the ranking, earn the click the system can’t discount. A result worth opening. A page worth staying on. A brand worth searching for by name. That’s the only version of this signal that survives a squashing function.

References

  1. USA v. Google LLC, trial transcript, testimony of Pandu Nayak, Google VP of Search. U.S. District Court for the District of Columbia. October 18, 2023.
  2. Baker, Barry Adams, et al. "What the Google antitrust trial revealed about ranking signals." Search Engine Land, 2023.
  3. Fishkin, Rand and Mike King. "An Anonymous Source Shared Thousands of Leaked Google Search API Documents." SparkToro, May 28, 2024.
  4. Schwartz, Barry. "Is CTR A Google Ranking Factor? Bartosz Góralewicz's Onely Experiment, With Comment From Google's Gary Illyes." Search Engine Land.
  5. Ahrefs. "AI Overviews Are Cutting Organic Clicks Even Further." Ahrefs Blog, December 2025 update.
  6. Federal Trade Commission. "FTC Announces Crackdown on Deceptive AI Claims and Schemes" (Operation AI Comply). Press release, September 25, 2024.
  7. American Bar Association. "Rule 7.1: Communication Concerning a Lawyer's Services." Model Rules of Professional Conduct.
  8. American Bar Association. "Rule 8.4: Misconduct." Model Rules of Professional Conduct.