How to Find Declining Pages in Google Search Console

The short version: open Search Console → Performance → Search results, turn on Compare in the date picker, and set a recent window next to an equal earlier one. Then click the Pages tab and sort by Clicks Difference — twice — so the biggest losers rise to the top. That sorted table is your list of declining pages. Export it, and you've got the pages quietly bleeding traffic, ranked by how much they've lost.

That's the whole method. Everything below is how to run it without misreading a normal dip as decay — and one honest rule for what actually counts as "declining."

I do this on my own sites by hand, so this is the manual version — no paid tool, just the free Search Console you already have. (If your site isn't verified in Search Console yet, Google's getting-started guide covers that first; everything below assumes you're in.)

Why declining pages hide

A page rarely dies loudly. It slips from position 3 to 6, loses a chunk of its clicks, and keeps sitting in your reports looking fine because it still gets some traffic. Your site total barely moves — a few winners cover for a slow leak across a dozen older posts. So nothing sets off an alarm, and the decay compounds for months before anyone looks.

Most advice on this stops at "check Search Console for drops." The pages that rank for it are either written for SEOs who already know their way around compare mode, or they're about diagnosing one big sitewide dip after it's already hurt. Neither gives you the routine: how to routinely surface the individual pages losing clicks, before the loss adds up. That's the gap this fills.

How to find declining pages in Google Search Console

Five steps. The whole thing hinges on two of them — turning on compare mode, and sorting by the difference rather than the raw total.

1. Open Search Console → Performance → Search results. This is the report with the four metric cards across the top: Clicks, Impressions, Average CTR, Average position. (Google's own reference for the Performance report defines each metric precisely, if you want the exact definitions.)
2. Turn on Compare in the date picker. Click the date filter, switch from the Date tab to the Compare tab, and pick two equal windows — a recent one and an earlier one to measure it against (the next section covers which windows to choose).
3. Switch to the Pages tab. Under the chart, click Pages. Now every row is one URL, with its clicks in each period side by side and the difference between them.
4. Sort by Clicks Difference — twice. Click the Clicks Difference column once (it sorts one way), then again, until the biggest losses sit at the top. This is the move: the largest negative numbers are your worst-decaying pages, ranked.
5. Export the list. Use the export button (top right) if you want to keep the ranked table in a sheet and track it month to month. Not required — but it turns a one-time look into a record.

At the end of this you have a list of pages ordered by clicks lost. That's the thing you were after — not a graph of a sitewide dip, but a ranked list of which pages are leaking.

Turn on compare mode (the setting most people miss)

Compare mode is the difference between "traffic feels down" and "these six pages lost these many clicks." Most people never leave the default Date view, so they stare at one line going down and can't tell which URLs are dragging it.

Here's the exact sequence:

1. Click the date picker (top-left of the Performance report).
2. Click the Compare tab.
3. Choose your ranges. "Compare last 28 days to previous period" is the quickest useful default. For a specific slump, pick Custom and set the recent range to the dip and the comparison range to a healthy stretch before it.
4. Keep Clicks and Average position switched on. Clicks tell you what you lost; position tells you why — a page that dropped clicks but held position has a different problem than one that fell down the results.

One more setting makes the comparison far easier to read, and most people miss it too. In the chart area there's a drop-down that says Daily — Google added weekly and monthly views there in December 2025. At daily granularity the weekends don't line up between your two periods, so the comparison chart turns into a sawtooth you have to squint through. Switch it to Weekly and the shape of the trend appears. Google's own note on the feature says it's "particularly powerful when comparing date ranges," which is precisely what you're doing here.

Once compare mode is on, the Pages, Queries, and Countries tabs all gain the difference columns. The Pages tab is where you find declining URLs; the Queries tab (filtered to one page) is where you later find which search term it lost.

Pick the right comparison window

This is the step that separates a real declining-pages list from a pile of false alarms — and it's the one the ranking articles skip. The window you compare against decides whether you're seeing decay or the calendar.

• For recent decay, compare 28 days against the prior 28. Equal length, same page. It's the cleanest read of "is this page slipping right now."
• For anything seasonal, compare year over year. Tax content falls off in May. Gift guides fall off in January. If you put December next to November, every seasonal page looks like it's dying. Switch the date picker to the 16-month view and compare the recent window to the same window a year earlier, so a normal seasonal dip doesn't read as decay. Sixteen months is the full window Search Console keeps, so that year-earlier comparison is available to you for as long as the page has existed.
• Ignore one-week wobbles. Clicks bounce week to week for reasons that have nothing to do with your content. A single quiet week isn't decay. A slide that holds across several weeks, on a page that used to earn steady traffic, is.

Get the window right and the Pages tab stops lying to you. Get it wrong and you'll spend an afternoon "fixing" a page that was always going to bounce back on its own next quarter.

Sort by Clicks Difference to surface the biggest losers

Raw clicks tell you which pages are big. The difference column tells you which pages are falling. You want the second one, so always sort by Clicks Difference, and always sort until the losses — not the gains — are on top.

Then read two numbers together on each row:

• Clicks down, impressions steady → the page is still being shown as often, but chosen less. Your position slipped just enough to cost click-through. Often this is a title and meta problem, not a content problem — the cheapest kind of fix, and worth treating as its own separate diagnosis before you rewrite anything.
• Clicks down, impressions also down → the page is surfacing for fewer searches. That's a ranking or relevance loss, and usually needs a real content refresh, not just a new title.

Same symptom on the surface — fewer clicks — two different causes and two different fixes. The impressions column is what tells them apart, which is why you keep it switched on.

What "decline" actually means

None of the top-ranking guides will tell you where normal fluctuation ends and decay begins, so here's the rule I use — plainly, so you can argue with it:

A declining page is one whose clicks have dropped roughly 20% or more below its ~90-day baseline, and that still earns at least about 10 clicks a week.

Two parts, both load-bearing:

• The ~20% below baseline keeps you from chasing every little wobble. A page has a normal range; decay is when it drops out of it and stays there. Comparing against a ~90-day baseline rather than last week is what makes "sustained" measurable.
• The ~10-clicks-a-week floor keeps small pages out of the list. A page that gets three clicks can "lose 67%" because two people didn't search this week. That's noise, not decay. Below the floor, the percentages are meaningless — so filter those pages out before you rank anything.

This is exactly the definition RefreshRadar runs on under the hood; I'm just showing you the manual version so you can do it yourself in the free tool.

A page that clears both bars is declining. Why it's declining is a separate read, and it comes down to what clicks, impressions and position did together — there are five patterns content decay makes, plus three impostors (an algorithm hit, a seasonal dip, a small-page wobble) that look identical in the clicks column.

From a list to a decision

A ranked list of declining pages is step one, not the finish line. Finding them is the easy part — Search Console did most of it. The harder questions come next:

• Which one do you fix first? Not the one that dropped the largest percentage — the one bleeding the most clicks per month, because that's the traffic actually at stake. If you want the formula for ranking the list by clicks at stake, that's covered in which blog posts to update first.
• Did the fix work? Once you refresh a page, you need to prove it recovered — same compare mode, one URL, before against after. That's its own method: how to tell if a content refresh actually worked.

Find, prioritize, verify. This article is the find. The other two close the loop.

Do this every week without doing it by hand

Running a compare-mode sweep once a month, sorting by clicks lost, filtering out the noise, and keeping the export up to date — it's a real routine, and it's exactly the kind of bookkeeping that quietly stops happening after the second busy week. That's not a knock on you; it's why decay wins. (If you'd rather keep doing it by hand, this sweep is check two of a five-check monthly Search Console routine that takes about fifteen minutes.)

That's the job RefreshRadar does for you. It watches every page in your Search Console continuously, applies the same decline rule (drop below baseline, above the clicks floor), and emails you the ranked list of pages losing traffic — each with the reason in plain English and the clicks/month at stake — so you never have to remember to run the sweep.

Connect Search Console read-only and see what's decaying on your site — free — no card, about 30 seconds, and we never post or change anything. If nothing on your site is decaying, we'll tell you that instead.