Google's 50-Week Search Console Bug: What Your SEO Data Is Really Telling You Now

Here's something that should make every SEO a little uncomfortable: for roughly 50 weeks — from May 13, 2025 all the way to April 27, 2026 — Google Search Console was quietly reporting inflated impression numbers to every single site on the web. Not a handful of sites. Not a niche bug. Every site. And most people either didn't notice, shrugged, or congratulated themselves on their "improved visibility."

Google confirmed the fix on May 4, 2026. The announcement was brief and clinical: "A logging error prevented Search Console from accurately reporting impressions from May 13, 2025 until April 27, 2026. This issue has been resolved." What they did not say — and what John Mueller later confirmed on Bluesky — is that the historical data will not be corrected. Those 50 weeks of inflated numbers are going to sit in your account forever. So if you're looking at year-over-year comparisons right now, you're comparing against a ghost.

This article is about what that means practically. Not just "your impressions dropped, don't panic" — but specifically which decisions you might have made based on bad data, how to figure out what your actual performance was, and how to build a clean baseline from this point forward. There's also a really underappreciated wrinkle about CTR and average position that I haven't seen covered anywhere properly yet, and we need to talk about that.

What Actually Happened — And Why It Matters

Let me be honest with you: Google has been vague about the technical specifics of what went wrong. We know it was a "logging error" in the impression tracking pipeline. We know it caused impressions to be over-reported. We don't know by how much — and that's actually the most frustrating part. Was it 10% inflation? 50%? 200%? Google hasn't said, and early reports from SEOs who noticed the fix going live suggest the drops have been significant. Some accounts saw impression counts fall by 30–60% almost overnight after April 27.

I've seen some people in SEO Twitter (or whatever we're calling it now) try to spin this as "no big deal since clicks weren't affected." I get why they're saying that — clicks are usually the more important business metric. But this framing misses something critical: impressions feed into two other metrics that SEOs and content teams actively use to make decisions. If your impressions were inflated, your click-through rate was deflated, and your average position data was skewed. Both of those numbers influence real strategic choices.

Think about the decisions that get made based on "my CTR is only 2% for this keyword." You write new title tags. You restructure your page. You add FAQ schema to try to get a featured snippet. You worry you're not competitive enough in the SERP. Now imagine you did all of that over the past year based on a CTR that was artificially suppressed because the denominator (impressions) was wrong. You might have been fixing a problem that didn't exist.

Or consider average position. A bunch of SEO content audits use average position as a signal for which articles are "stuck" on page 2 vs ranking solidly on page 1. If the position data was being pulled from the same broken logging pipeline — and it was, since GSC derives it from impression events — then your position data for the affected period is unreliable too. Not completely useless, but definitely not trustworthy for fine-grained analysis.

Which Metrics Were Poisoned (And Which Were Fine)

Let's get specific about this, because I've seen a lot of confused takes floating around. Google was clear that clicks were not affected. The logging error was specifically in the impression counting system. So your actual traffic numbers — the people who physically clicked through to your site — those are accurate. Your clicks trend data from the past year is fine.

The Impression Cascade Problem

Here's what wasn't fine. Search Console calculates CTR as clicks divided by impressions. If impressions were inflated, then your CTR for every single query in the affected period is lower than reality. The math is simple: if you got 100 clicks from what was logged as 5,000 impressions but the real number was 3,000 impressions, your real CTR was 3.3% — but GSC showed you 2.0%. That's a massive difference in how you'd evaluate a piece of content.

Average position is even trickier. GSC calculates it as a weighted average across impression events. More impressions at a certain position shifts the average. With artificial inflation in the impression data, the position averages were being calculated on a distorted input. In practice, I suspect position data was less severely impacted than CTR data (since the inflation likely affected impressions at all positions somewhat proportionally), but you should still treat any position-based analysis from that period with healthy skepticism.

The metrics that are totally clean: clicks, sessions in GA4, conversions, and any server-side data you collected. If you were also running Bing Webmaster Tools, that data was completely separate and unaffected. Same for any rank tracking tools you use — those pull live SERP data independently, not from GSC's logging system.

What About Discover and Google News?

Google's announcement specifically mentioned the Performance report for Search results. The Discover and Google News reports in Search Console have their own separate impression tracking pipelines, and Google was not explicit about whether those were affected. Based on the wording ("Search Console Performance report") the bug appears to be web search specific, but if your site gets significant Discover traffic, I'd approach those impression figures from the same time window with caution until we get more clarity. The safe play is to treat clicks as your source of truth across all report types for that period.

How to Audit Your Historical Data Right Now

Okay, so the data is broken for that window. What do you actually do about it? First, don't delete anything. GSC doesn't let you export and reimport anyway, but more importantly, you still have your clicks data which is valuable for trend analysis. The goal here isn't to throw out the period — it's to understand what you can trust and what you need to flag.

The audit process I'd recommend has four phases. You're going to cross-reference your GSC clicks data with GA4 organic sessions to establish a reliable performance picture, identify any decisions you made based on the bad impression/CTR data, assess whether those decisions were probably still correct (many of them will be), and then set a clean baseline from April 28, 2026 onward.

The Cross-Reference Method

Here's the practical technique: go into Google Search Console and export your clicks data by page for the full affected window (May 13, 2025 – April 27, 2026). Then pull organic sessions by landing page from GA4 for the same period. These two numbers should broadly correlate — GSC clicks are not exactly the same as GA4 sessions (there's always some discrepancy due to JavaScript loading, bot filtering, etc.) but typically they move together. If you see any major divergences on specific pages, that's worth investigating.

What you're really looking for is whether your actual traffic trend (clicks and sessions) was growing, stable, or declining — and using that to cross-check any strategy decisions you made based on GSC impression or CTR data. Most sites I've seen discussed in the community found their clicks trend was actually fine throughout the period. The bug was inflating impressions without meaningfully affecting the click-to-site rate in practice. But you want to verify this for your own site rather than assume.

• Export all GSC data from May 13, 2025 – April 27, 2026 (Clicks column is trustworthy)
• Pull GA4 organic landing page data for the same date range and compare the trends
• Flag any pages where you rewrote title tags or meta descriptions based on "low CTR" signals during this period
• Review any content audits that used average position as a primary ranking signal — those need revisiting
• Check if your reporting dashboards have historical impressions charts that will mislead stakeholders after the drop
• Annotate your GSC account (or your reporting tool) to mark May 13, 2025 – April 27, 2026 as "impression data unreliable"
• Document what a "normal" impressions range looks like for your site from clean post-April-28 data

Rebuilding an Accurate Performance Baseline

This is the part nobody is talking about enough. Yes, the old data was bad — but the more pressing question is: what does "normal" even look like for your site now? Because starting April 28, 2026, your impressions number probably dropped significantly. If you go into GSC today and compare this week to any week from the last year, you're going to see a scary-looking decline. Your team is going to ask questions. Your clients are going to ask questions. You need to be ready with a clear answer.

The first thing I'd do is take a 30-day snapshot of your impressions, clicks, CTR, and average position from April 28 onward. This is your new baseline. For most sites, your clicks should be relatively stable compared to the pre-fix period (because clicks weren't affected). What you'll notice is that your CTR probably looks much better now — because the same number of clicks is now being divided by a more accurate (lower) impressions figure. This is a good thing. Your site's actual click-through performance was probably better than you thought all along.

Average position is worth watching too. For many sites, position data actually becomes more stable and reliable after the fix. The pre-fix inflation likely added noise to position calculations by injecting artificial impression events. Your position data going forward should give you a cleaner read on where you're actually ranking for specific queries.

One practical thing you can do right now: use the date comparison feature in GSC to compare April 28 – May 9, 2026 against the same date range in 2024 (which was before the bug started). This gives you a legitimate year-over-year window that isn't contaminated on either side. Your clicks trend from that comparison will tell you the real story about whether your organic performance has grown or declined over the past year.

Re-Evaluating Title Tag and Meta Description Decisions

Here's the uncomfortable one. If you spent any part of the last year iterating on title tags or meta descriptions to improve CTR — because your CTR data was showing you below-average numbers — you need to go back and reassess whether those changes were actually necessary or beneficial.

That doesn't mean your new title tags are wrong. They might be genuinely better. But if the entire premise for the change was "our CTR for this page is 1.8% and that's below average," you should know that the 1.8% was a fiction. The real CTR might have been 2.8% or 3.2%, which could have been perfectly competitive. Run a clicks-based analysis on those pages to see whether click volume actually improved after your title tag changes. That's the real signal.

The same logic applies to any pages you might have flagged for content rewrites because they showed "high impressions, low clicks" patterns. That pattern can be a real signal — it often points to keyword intent mismatches — but it can also just be inflated impression noise. Look at your actual engagement metrics (GA4 bounce rate equivalent, time on page, scroll depth) to figure out which category each flagged page falls into.

What to Do Going Forward

The most important shift you can make after this is to diversify your SEO measurement. Search Console is a fantastic tool and I'm not suggesting you abandon it — but this episode is a good reminder that any single data source can have problems, and the SEOs who got burned the most by this bug were the ones who were making decisions purely based on GSC impression and CTR data without cross-checking against other signals.

Going forward, clicks should be your primary GSC metric. Clicks are robust, they weren't affected by this bug, and they directly measure what actually matters — people actually navigating to your site. Impressions are useful for broad reach analysis but shouldn't drive granular content decisions. CTR is fine as a directional signal, but always triangulate it against actual page analytics before rewriting anything.

I'd also strongly encourage using independent rank tracking for your most important keywords rather than relying solely on GSC average position. GSC gives you a great overview across thousands of queries, but for your core 20-50 keywords, a dedicated tracker that hits the SERPs directly gives you much more precise and reliable position data. Combined with a proper keyword volume check for your target terms, you get a much more complete picture than GSC alone provides.

Here's the other thing I want to say clearly: don't let this bug shake your confidence in SEO as a discipline. The fundamentals — create useful content, earn genuine authority, keep your site technically clean, build real topical depth — none of that changed. Your site's actual organic performance over the past year was driven by those factors. The broken GSC data just made it harder to see clearly. Now the data is fixed, and if you approach the next few months with clean measurement practices, you'll have a much better foundation for making smart decisions.

Start fresh. Set your new baseline. Document what "normal" looks like for your site now. And next time someone suggests making major content decisions based on a single metric from a single data source, push back. Good SEO measurement has always required triangulating across multiple signals — this bug just made that lesson a little more expensive for the people who didn't already know it.