How to Verify Rank Tracker Data After Reports of Google Poisoning SERP Results

To verify rank tracker data, compare it with at least three other sources before you report a change: Google Search Console, a manual logged-out check and your own traffic data. Ahrefs said on October 22, 2026 that between 5 and 20 percent of search results are affected by altered results shown to data providers.

This guide is for B2B SaaS marketers and founders who report on their own rankings. Cekpr covers SEO, AI search and content marketing for B2B software, and this piece explains what the reports say, what is still unknown and a simple check you can add to your monthly review so a bad data point does not become a bad decision.

What Did Ahrefs Report About Google and SERP Data?

Ahrefs said Google is showing different results to suspected SERP data providers. In its October 22, 2026 product blog post, the company called this "poisoning," described it as an industry-wide problem and said that "between 5% and 20% of search results" are affected, depending on the data provider.

The post lists what Ahrefs is doing about it. It runs filters to reject abnormal results and request accurate replacements, it is cleaning up past results that may have been affected, and it is monitoring the situation. It also warns that what works to get accurate data today may not work tomorrow.

The post gives no sample size, start date or method. That is a limit worth stating plainly. The claim comes from a company that sells rank data, so treat the exact percentages as a signal, not a measurement you can audit.

What Do Other Tool Makers Say?

Other vendors report different numbers, and the spread is wide. PPC Land's summary of the story collects them: DemandSphere reported about 11 percent of daily rank checks on October 13, Nozzle reported 8.0 percent of pages on October 19, and AccuRanker reported under 1 percent in mid-October.

Nozzle also gave timing. According to the PPC Land report, it saw the first poisoned pages on September 20, with a peak of 27.0 percent on September 29. AccuRanker dated the onset to September 24, with smaller rounds of swapped pages from January 2026.

Those figures are not directly comparable. Each vendor counts different things, such as results, pages or daily checks. What they share is the direction. Several independent tools saw odd results, and none reported a clean record.

Why Does This Matter for a B2B SaaS Team?

It matters because rank reports drive decisions about budgets, content and agency contracts. A fake drop can trigger a panicked rewrite. A fake win can hide a real problem. And a SaaS site with a few hundred tracked keywords can have its whole report skewed by a handful of bad rows.

The timing makes it harder. Google's September 2026 spam update, which PPC Land dates from September 24 to October 8, overlapped with the window when vendors saw the most altered data. Real ranking changes and bad data may sit in the same weeks. Our backlink audit after the September 2026 spam update covers the real-effects side of that window.

Cekpr's view is simple. Do not throw out your tracker, but do not let it be the only voice in the room.

How Do You Check Rankings With Four Sources?

Use four sources and look for agreement. A rank change is credible if at least three of them point the same way. If only the tracker moves, treat it as unconfirmed until another source catches up.

Use this checklist for each keyword group that changed.

The order matters little. What matters is that a single tool does not get to declare victory or disaster. If the manual check shows your page where the tracker says it vanished, the tracker is the odd one out.

How Should You Read Search Console as a Cross-Check?

Search Console is the best cross-check because the data comes from Google itself, and not from a scraper. Google's Performance report documentation says the report shows clicks, impressions, click-through rate and average position, and that the newest data can be preliminary and may change within hours.

Average position needs care. Google explains that the chart shows the average position of the topmost result from your entire site, while the table shows the average for the specific page or query in that row. A site that ranks for more queries can see its average position move without any single ranking changing.

So lean on clicks and impressions for the page-level story. If a landing page lost clicks and impressions together, the drop is probably real. If rank fell in the tracker but clicks held steady, suspect the tracker.

What About Manual Spot Checks?

Manual checks are slow and imperfect, but they answer a narrow question well: what does a real searcher see? Search results vary by location, device, language and personalization, so write down the conditions each time.

Keep the sample small and consistent. Pick ten queries that matter to the business, check them in a logged-out window on the same device and compare to last month. You are looking for gross mismatches, not a one-position wobble.

If your SaaS targets specific countries, check from those countries using a location setting or a trusted colleague. A US-only check can miss a drop that affects Canada or the UK.

What If the Four Sources Disagree?

When sources disagree, trust the one closest to real users and wait for a second week of data. Search Console and your own analytics reflect actual searchers and visitors. A rank tracker reflects what its crawler was shown, which is the part now in question.

Mixed results are normal. A tracker may show a drop while Search Console shows flat clicks, which usually means the tracker saw an odd page. Or Search Console may show lower impressions while the tracker looks fine, which can mean demand fell, not rank. Name the pattern and move on.

Do not make big content or budget moves on a mixed signal. Give it a week, re-check the same queries and see whether the sources converge.

How Do You Annotate and Report the Affected Weeks?

Mark September 20 through at least mid-October in every chart that uses third-party rank data, and say why. A short note such as "tool vendors reported altered SERP data in this period" tells readers to treat that stretch carefully. It also protects you if the data is later revised.

Ahrefs itself says it is cleaning up past results, so a chart exported in early October may not match the same chart viewed later. Save a dated copy of any report you plan to cite. Without it, you cannot explain why last month's number changed.

Then add a confidence label. For each reported change, write confirmed, mixed or unconfirmed. Leaders do not need the full story, but they do need to know which numbers to lean on.

What Should You Change in Your Monthly Review?

Add a data-quality step to the review, and keep it light. Ten minutes of cross-checking before the meeting prevents an hour of arguing about a number nobody trusts. It fits naturally into a regular content audit, and our quarterly and annual content review schedule shows where to place it.

If traffic really did drop, the data check is only the first step. Our guide to diagnosing a SaaS organic traffic drop walks through the five usual causes once you are sure the drop is real. And if you want more first-party data on visibility, see how to measure AI search visibility in Search Console.

Finally, ask your vendor. Send a short question: how are you detecting abnormal results, how far back does your cleanup go and will you flag corrected data? Good vendors answer plainly. The reply is useful evidence in itself.

Conclusion

The reports of altered SERP data do not mean your rank tracker is useless. They mean it needs a second opinion. Compare it with Search Console, manual checks and your own traffic before you act on a change.

Cekpr will keep following how vendors and Google respond. Until there is clearer evidence, annotate the weeks since September 20, label your confidence and let a rank change count only when more than one source agrees.