One SEO metric I think deserves more attention: the percentage of published URLs that get at least one organic click over a set period. Most dashboards focus on total indexed pages, impressions and clicks. Useful, but they can hide a growing tail of pages that never attract real search demand. A site can publish thousands of new URLs while nearly all traffic still comes from a small group of older pages. I like looking at it by cohort: group pages by publication month and template, then track for each cohort the share of URLs with at least one click in 90 days, impressions per URL, index status and how many queries each page ranks for. Important caveat: there's no universal safe ratio. New pages need time, niche pages can be valuable with low volume, and some pages exist for users rather than search engines. It's a diagnostic, not a known ranking factor and not a reliable predictor of an update on its own. Still, a steadily falling share of pages that get clicks can be an early sign that publishing is outpacing demand or differentiation. Has anyone tracked something like this across a few core updates? Did it show problems earlier than total organic traffic did?
Haven’t tracked it across updates, my sites are way too small for cohorts to mean much. On a small site I’d rather just look at the list of URLs with zero clicks in GSC once a quarter, it’s basically the same signal without the dashboard.
Fair point. Below a few hundred URLs the cohorts get too thin to say much, and a quarterly zero-click list from GSC does the same job. One small addition: sort that list by publish date. A zero-click page from last month mostly needs time, but one that has been live for a year without a single click is the one to improve, merge into a stronger page, or retire.
I like this metric because total traffic hides decay. A few top pages can carry the average while the long tail dies. Two practical tips if anyone builds it: - Use the Search Console API or the BigQuery bulk export, not the UI. The UI's 1,000-row limit silently drops long-tail URLs, which is exactly the cohort you're measuring. - Exclude pages younger than ~90 days, or each new batch will drag the ratio down and look like a problem. Plotting "% of URLs with at least 1 click" per template against core-update dates should make it obvious whether one template type is being hit, before it shows up in total sessions.
One thing that helps with this metric is to exclude pages that are meant to get no organic clicks, like thank-you pages, tag archives and pagination, before you calculate the share. Otherwise a healthy site can look worse than it is. I would also compare cohorts of the same template only, since a product page and a blog post have very different click patterns.
One useful split I would add: track the funnel in two stages, published-to-indexed and indexed-to-clicked, rather than published-to-clicked alone. A falling share of pages with clicks can mean two very different things: the pages are indexed but losing rankings, or they were never indexed in the first place and could not possibly get clicks. Pull the index coverage data alongside the Search Console clicks for the same cohort, and the gap tells you whether you are dealing with a crawl and indexation problem (fix internal links, sitemaps, crawl budget) or a content problem (consolidate or improve the zero-click URLs). Comparing each cohort against its own previous month also controls for seasonality, which raw totals will happily mislead you about.
I track this in a simpler way on a small content site, and it lines up with what siteproofhq said about excluding young pages. I pull the Search Console list once a quarter, filter out anything published in the last 90 days, and then look at the zero-click list by template — in my case, item pages versus guides. The item pages almost always have a long tail of zero-click URLs, but many of them are still worth keeping because they answer a specific query and support the pages that do rank. What I found useful was splitting the zero-click list the way rbsojib suggested, into indexed-but-no-clicks and not-indexed-at-all. The first group gets a title and intro refresh, and the second group usually has a real reason — thin content, duplicate intent with another page, or no internal links pointing to it. That split keeps me from merging or deleting pages that just need a better title.
Honestly for small local-business sites I'd flip it and track the handful of pages that actually bring in calls or form fills, usually the service pages and the location pages, instead of the share of all URLs. Most of those sites only have 10 to 30 pages, so a page with zero clicks is often just a thin 'about' or blog post, not a signal that something's wrong. What I'd check in Search Console is whether each service + city page gets impressions for its own terms, and if it doesn't, that page needs a rewrite or a few internal links before any new content gets added.
Thanks, a few of these improve the metric more than my original version did. rbsojib's two-stage split (published to indexed, then indexed to clicked) is the one I'd adopt first, because it tells you which problem you have before you touch any content. siteproofhq's point about the 1,000-row limit in the UI matters for the same reason: the long tail is exactly the part that gets cut off. NinjaSitePartner, fair for a 20-page local site. A share of URLs says little there, and checking whether each service page gets impressions for its own terms is the better test.
The two-stage split makes this much more useful in practice. A page that was never indexed and a page that is indexed but never clicked are different problems, and lumping them together hides which one to fix first. One extra split I would add is by intent. Informational pages and commercial pages should not share the same target share — a blog post can be doing its job by earning internal-link traffic and supporting a service page even if it rarely gets an organic click itself. Tracking them together makes informational content look like dead weight when it is not.
I like treating this as a diagnostic rather than a ranking target. For a small product site, I would keep the check simple: group pages by type and publication month, then compare their age, impressions and clicks after they have had time to settle. A new listing with no clicks is not the same as an older page that has never found search demand. I would also keep useful customer-support pages in the mix even if their search numbers are modest. That makes the review practical without turning every page into a traffic target. Do you find grouping by template gives a clearer signal than grouping by topic?
@TUITT by template, in my experience. Topic groups mix page types with very different click expectations, so a weak topic often turns out to be one weak template inside it. Grouping by template first shows whether the template itself works (title pattern, internal links, how unique each page really is), and topic becomes the second cut once you know the template is fine. Your point about age matters for that too. I'd only count a page once it's past a settling window, otherwise every fresh batch drags the share down for a few weeks and looks like a problem it isn't. @Steven Mock agreed on the intent split. Informational pages judged on organic clicks alone will always look like dead weight, even when they're the ones feeding the service pages.