Google SERP Poisoning: Is Your Search Console Data Real?

Your Search Console impressions and average position are the least trustworthy numbers in your SEO report, and your agency shouldn't lead with either of them. Google is fighting the scrapers, rank trackers and AI bots that hammer its results pages, and that fight leaves fingerprints in your data. Some automated requests show up as impressions. Some trackers get served results that don't match what your customers see. Clicks, organic sessions, conversions and revenue are much harder to fake. So judge your SEO, and your agency, on those. Read impressions as a rough signal of visibility and nothing more.

The rest of this piece explains why that holds, how to test your own account in about an hour, and what to ask the people reporting to you.

What SERP poisoning means for the data you see

"SERP poisoning" is the name for Google's defenses against automated access to search results. Those defenses include blocking scrapers, forcing heavier rendering, and serving altered or degraded results to traffic Google suspects isn't human. The goal is to make large-scale scraping expensive and unreliable for the tools that resell rankings and for AI systems that harvest answers.

We don't have visibility into Google's bot detection. Google hasn't published how much automated traffic Search Console filters out before it reports impressions. So we won't put a percentage on the inflation in your account, and you should be wary of anyone who does.

What we can do is describe the mechanism. It's enough to change how you read your report.

Impressions count when your result appears on a loaded results page. A person doesn't have to look at it or scroll to it. If an automated request loads a results page and Google doesn't filter that request, your site can log an impression that no customer produced.

Deep results inflate impressions and drag down average position. Rank tracking tools have historically pulled many results per query so they can find a site sitting at position 47. If a tool loads results that deep and your page appears at 47, that can register as an impression at position 47. Repeat that across thousands of tracked keywords and you get a pile of impressions at positions no human visits. When Google blocks those requests, your impressions drop and your average position "improves," and nothing about your actual visibility has changed.

Poisoned trackers report rankings your customers never see. If Google serves a tracker an altered page, the rank it records is a rank for a bot. Your agency's ranking dashboard can be wrong in either direction.

Clicks are the hardest metric to fake through scraping. A scraper wants the page. It doesn't need to click through to your site. That's why clicks hold up better than impressions when automated traffic rises or falls.

That's the Google SERP poisoning Search Console data accuracy problem in one sentence. The metrics that move most when bot traffic changes are the two metrics agencies love to put on slide one.

Why rank still matters, even when rank data is noisy

You could read all this and decide rankings don't matter anymore. That would be a mistake.

In a July to September 2026 analysis of client sites we manage, we looked at 2,350 tracked keyword snapshots where a Google AI Overview appeared, covering 314 keywords. The AI Overview cited the site 73% of the time when the site held organic position 1. That fell to 49% at positions 2 to 3, 42% at positions 4 to 10, 9% at positions 11 to 20, and 3% when the site ranked beyond position 20 or didn't rank at all. The pattern held on nearly 80% of the sites. The rest had too little data to judge. It's a correlation, not proof of cause, and we wrote up the full findings in AI Overview Citations Follow Rank: What Our Data Shows.

The point for this article is simple. Position is worth knowing, because it tracks with whether AI Overviews cite you. That makes accurate rank data more valuable, not less. A tracker that a poisoned results page fooled can tell you you're at position 3 when you're at 12, and on our numbers that's the difference between a 49% citation rate and a 9% one.

One more detail from that analysis matters here. About a quarter of citations went to a different page on the same site than the page that ranked. If your reporting only tracks one URL per keyword, it misses part of your visibility no matter how clean the data is.

Seven checks you can run on your own Search Console this week

None of these needs a tool beyond Search Console, GA4 and a browser. Set the date range to the last 16 months so you can see the shape of things over time.

1. Plot clicks and impressions on the same chart

Turn on both lines in the Performance report. In a healthy account they move roughly together. When impressions climb hard while clicks stay flat, CTR falls, and someone usually blames AI Overviews. Sometimes that's right. Sometimes the impressions were never human.

When impressions drop sharply in a single week with no matching drop in clicks, that's the more telling pattern. Real visibility losses take clicks down with them. A sudden impression cliff with flat clicks points to automated traffic leaving the count.

What to do: Mark any week where impressions moved more than clicks by a wide margin. Compare those dates against your own site changes. If you didn't ship anything that week, treat the impression change as a data event, not a performance event.

2. Split by device

Scraping and tracking tools have long favored desktop requests. Filter the Performance report by device and compare the desktop impression line against mobile.

If desktop impressions spiked or collapsed while mobile stayed steady, and desktop clicks didn't follow, that's a strong signal the desktop movement came from bots.

What to do: For an e-commerce store whose buyers shop on phones, weight mobile clicks more heavily in your read of SEO health. Don't let a desktop impression swing drive a strategy decision.

3. Stop reading sitewide average position

Sitewide average position blends every query you show up for, including thousands at positions 40 to 90 that no buyer reaches. One change in how bots get counted can move that number by several places overnight.

What to do: Filter to queries where you rank in the top 20 and sort by clicks. Track average position for your top 25 to 50 revenue queries individually. A move from 6 to 3 on a query that drives sales matters. A sitewide shift from 28 to 22 tells you almost nothing.

4. Find the queries with impressions and zero clicks

Export the Queries report for the last 90 days. Sort by impressions and filter for zero clicks. Look at what's at the top.

Some of it will be real. Brand-adjacent searches answered in the results, informational queries an AI Overview handled. But watch for odd, overly specific strings, long lists of near-identical variants, or queries with no plausible human intent. Those are the fingerprints of tools querying at scale.

What to do: Don't build content plans off queries that have never produced a click. Prioritize queries that have clicks and convert in GA4.

5. Check the country filter

Filter by country. If you sell only in the US and a large share of impressions come from markets you don't serve, with almost no clicks, ask why.

What to do: Report SEO performance for the markets where you actually take orders. Everything else is noise for a business decision.

6. Reconcile Search Console clicks against GA4 organic sessions

Search Console clicks and GA4 organic sessions will never match exactly. They count different things, and consent banners and tracking gaps widen the difference. What should stay fairly stable is the ratio between them over time.

If the ratio shifts suddenly, you have a measurement problem somewhere. It could be in Search Console, in your tag setup, or in how AI referral traffic gets classified. We covered the tracking side in ChatGPT Traffic Showing as Direct in GA4: Audit & Fix.

What to do: Track the clicks-to-sessions ratio monthly. Investigate any month where it moves noticeably, before you trust either number.

7. Spot-check your tracker with a manual search

Pick ten revenue keywords. Search each one from a clean browser session on your phone, in the city where your customers are. Write down where you actually appear. Compare that to your rank tracker and to Search Console's average position for the same query.

Personalization and location mean you won't get perfect agreement. You're looking for big gaps. If your tracker says position 2 and you can't find yourself on page one from three different devices, the tracker's data is suspect.

What to do: Run this check quarterly. Ask your agency to run it with you on a call. It takes 15 minutes and tells you a lot about whose data they trust.

Use paid search as a second read on demand

If you run Google Ads, you already own a second dataset on what people search for. The search terms report shows the actual queries that triggered your ads and got clicked. Paid clicks cost money, which gives Google a strong reason to filter invalid activity out of that billing data.

Compare the organic queries Search Console says matter against the search terms that drive paid conversions. When a query shows heavy organic impressions but never appears in your paid search terms, despite your bidding on that theme, question whether that demand is human. When the two sets line up, you can trust the organic impression trend more.

Running SEO and paid search under one roof makes this comparison routine instead of a once-a-year project. If your paid account isn't set up to give you a clean query read, our google ads management services are built to tie spend, search terms and organic visibility together in one view.

How to judge whether your agency's numbers are real

You don't need to audit Google's bot filtering to judge your agency. You need to look at what they choose to put in front of you, and how they react when you question it.

What a report should lead with

Organic revenue and conversions for the markets you sell in come first. Organic clicks and sessions, broken out by device, come next. Then position for a named list of revenue queries, checked against reality. Impressions belong further down, labeled for what they are, with a note on any week where they moved without clicks.

Red flags

• Impressions on the first slide. If the headline win is "impressions up," ask what happened to clicks and revenue in the same period.
• Sitewide average position as a KPI. It's too easy to move without doing anything.
• Rank tracker screenshots with no manual verification. Ask when they last checked a tracked keyword by hand.
• No device or country split. Blended data hides exactly the patterns bot traffic creates.
• Silence about measurement changes. When impressions drop off a cliff and the agency explains it as an algorithm update without checking clicks, they're guessing.
• Every month is up. Real SEO has flat months and bad months. A report with nothing but green arrows is selecting metrics.

Questions to ask on your next call

1. "Show me organic revenue for the last 12 months next to organic impressions. Where do they diverge, and why?"
2. "Which 25 queries drive our organic revenue, and where do we rank for each one today, verified by hand?"
3. "What's our desktop versus mobile split on impressions and on clicks?"
4. "When did you last reconcile Search Console clicks against GA4 sessions, and what was the ratio?"
5. "If Google changed how it counts bot traffic next month, which of our KPIs would move, and which wouldn't?"

That last question is the best test. An agency that understands its data can answer it in one sentence: impressions and average position would move, clicks and revenue wouldn't. An agency that can't answer it is reporting numbers it doesn't understand.

Why we report this way

In Metrix Digital's earlier years, building monthly client reports in Looker Studio was part of Michael Pam's own job. Clients glanced at the trend lines and moved on. That taught us something about the trend lines themselves. A chart full of rising impressions looks like progress and asks nothing of the reader, which is exactly why it's dangerous when the impressions aren't real. We wrote more about this in Your Monthly Report Is Dead. Your Client Never Opened It Anyway.

Today we run client marketing through Athena, our own system, and we keep static records of what ran as proof. When a metric moves, we want to point to the work that moved it, or say plainly that it moved because of a measurement change on Google's side.

That discipline comes from owning the whole stack. We do strategy, SEO, design and Magento development under one roof, so when organic clicks and on-site revenue disagree, we can check the tracking, the templates and the search data ourselves. Nobody hands off the problem to another vendor. For WaterBottles.com, our work spans e-commerce strategy, the Magento build and SEO. For CarWraps.com, it's local search marketing, web design and lead generation. In both cases the same team that reads the search data also owns the site that has to convert it. The proof we stand behind is retention. Clients stay with us for years, and they don't stay for impression charts.

Where we could be wrong

If Google filters automated traffic out of Search Console far more completely than its outside behavior suggests, your impressions are cleaner than we're implying, and impression trends deserve more weight. What would change our answer is either Google publishing detail on how it filters bot traffic from Search Console, or your own account showing impressions and clicks moving in lockstep across devices and countries for a year or more. If you see that second pattern, trust your impressions more. Until then, treat them as a visibility signal, and run the business on clicks and revenue.

The short version

• Impressions and sitewide average position are the metrics most exposed to bot traffic and Google's countermeasures. Don't make decisions on them alone.
• Clicks, organic sessions, conversions and revenue are much harder for scrapers to distort. Make them your headline.
• Rank still matters. In our data, AI Overview citations track closely with organic position, so verify your rankings by hand instead of trusting a tracker blindly.
• Split by device and country, check zero-click queries, and reconcile Search Console against GA4 every month.
• Ask your agency what would move if Google changed bot counting tomorrow. Their answer tells you whether their numbers are real.

If you want a second opinion on your own reporting, we'll run these checks against your Search Console, GA4 and paid search data and show you which of your numbers hold up. Get an audit, or request a proposal if you're ready to have one team own the result from data to launch.