Impressions Went Up. Position Got Worse. Which One Do You Believe?
When a page's search impressions climb while its average position worsens over multiple readings, the page usually isn't ranking better, it's being matched to a wider set of queries it doesn't actually answer well. Check whether the newly-surfacing queries match the page's actual intent before treating rising impressions as a win.
We caught this on one of our own pages. Impressions kept climbing, reading after reading, and it took three of those readings before we stopped reading that as good news.
The pattern that looks like growth
A glossary-style page on our own site logged four consecutive weekly readings. Impressions: 8, then 11, then 17, then 27, climbing every single time, which is the number most dashboards surface first and the one that's easiest to screenshot into a status update. Average position over the same four readings: 15.4, then 17.3, then 28.8, then 23.0. Worse, worse, worse, then a partial recovery on the fourth read that's one data point, not yet a trend.
Read the impressions line alone and it looks like a page finding its audience. Read the position line alongside it and a different story shows up: the page is getting shown to more searches, and ranking worse on average across that larger set, for three straight readings before any recovery. Those aren't two separate facts. They're the same underlying event described from two angles.
Why this happens: you're being matched broader, not ranking higher
Google Search Console's impression and position metrics answer different questions. Impressions count how often a page appeared in search results for any query, regardless of where. Average position is the mean rank across all of those appearances. A page can gain impressions purely by getting swept into more query matches, including ones only loosely related to what it actually covers, while its ranking on the queries it was actually built for stays flat or slides.
This is a structural risk for any page with broad, glossary-style content, because that's exactly the content shape search engines match loosely: define enough terms and you become a plausible partial answer to a lot of adjacent searches you never intended to target. The page isn't getting worse at its job. It's being asked a wider range of questions, most of which it wasn't written to answer well, and its average position reports the honest result of that mismatch.
The alternative explanation, real ranking improvement, would show the opposite shape: impressions and position moving together, or position holding steady while impressions grow within the terms you're actually targeting. When the two move in opposite directions for multiple consecutive readings, broader matching is the more likely read, not a coincidence worth ignoring.
The check: look at what's actually matching before you react
Before treating either number as the story, pull the query-level breakdown, not just the page-level summary. Three questions settle which pattern you have:
Are the newly-appearing queries close to what the page is actually about, or adjacent-but-different? A glossary page picking up impressions for the exact terms it defines is healthy growth. The same page picking up impressions for tangentially related searches it happens to mention once is the broad-match pattern, and no amount of impression growth there converts to relevant traffic.
Is click-through rate on the page flat or falling as impressions rise? If impressions are up but clicks aren't following proportionally, searchers are seeing the page and not finding it worth a click for the specific thing they searched, a signature of broad, low-relevance matching rather than real demand growth.
Has this happened on one reading, or several in a row? One reading of impressions-up-position-down is noise; SERP volatility alone can produce it. Three or four consecutive readings in the same direction is a pattern worth acting on, not a fluctuation worth waiting out.
If the answer to the first question is "mostly adjacent, not on-target," the fix isn't to abandon the page. It's usually to narrow or split the content so it stops trying to be a plausible partial match for everything nearby, and instead earns a strong match for the terms it was actually built to own.
Information gain: why this matters more for a low-authority property
This pattern is easy to miss on an established site with a deep backlink profile and hundreds of ranking pages, because a handful of pages drifting into broad-match territory barely moves the aggregate numbers anyone's watching. On a newer or lower-authority property, where total indexed queries can be counted on one hand for weeks at a time, this exact pattern can dominate the entire search visibility picture. One page finding broad matches looks, from a distance, like the whole site is gaining traction. It isn't. It's one page's relevance getting diluted, and because there's little else to average it against, the dilution is loud instead of quiet.
That's the specific reason this is worth checking rather than assuming: the smaller the site, the more a single page's broad-match drift can masquerade as overall growth, and the more a founder relying on a weekly glance at "impressions are up" can walk away with exactly the wrong read.
What to do with the read once you have it
If query-level data confirms broad, low-relevance matching, don't treat it as a content failure to fix immediately. Treat it as a data point about what the page is actually being matched against right now, log it, and watch whether it self-corrects as the page accumulates more crawl history, or whether it holds across enough readings to justify a content split. Reacting to one reading, in either direction, is the same mistake the constellation-not-ladder approach warns against in reconciling conflicting signals: one number doesn't decide anything on its own, and a pattern held across multiple independent readings does.
FAQ: My search impressions are rising but my average position is getting worse. Is that good or bad?
Neither, on its own, until you check query-level detail. Rising impressions with worsening position across multiple consecutive readings usually means the page is being matched to a broader, less targeted set of queries rather than ranking higher for the ones it was written to win. Pull the actual matching queries, check click-through rate against the impression trend, and confirm the pattern holds across at least two or three readings before deciding whether to act. A single reading either way is closer to noise than signal.
Unfamiliar with a term above? The P8 glossary defines P8's own vocabulary without the jargon, the same kind of page this pattern was caught on, in fact.
Founders without a dedicated SEO function tend to hit this exact ambiguity when they don't have the query-level habit built in yet. P8 for founders covers what usually fills that gap.
See how P8 attaches the underlying evidence to every recommendation, so a number like this comes with the context needed to read it correctly. Start with your first brief.
Figures in this article reflect P8's own Google Search Console readings across four consecutive weekly pulls. Product capabilities described reflect P8's current approach, not a guarantee of specific features.
