Leads from the site fell by a third last quarter. The design team believes the page is confusing. The SEO team believes rankings slipped. Both explanations account for the chart, and the company is about to spend money on whichever team argues more persuasively.
That argument is avoidable. A visit passes through three layers, each one measured separately, and the numbers will say which layer failed before anyone redesigns anything.
What are the three layers a visit passes through?
Getting shown, getting clicked, and getting used. Impressions measure the first, click-through rate the second, and on-page behavior the third. A failure in one leaves the other two intact, which is what makes it identifiable.
Layer one is search: does your page appear in organic search for the query at all. Layer two is presentation: given that it appeared, did anyone choose it over the results around it. Layer three is the page: given that they arrived, did they get what they came for. Only the third is a user experience problem in the sense a designer means.
| Impressions | Clicks / CTR | On-page behavior | Where the problem is |
|---|---|---|---|
| Down | CTR steady | Steady | Search — ranking, indexing or demand |
| Steady | CTR down | Steady | Presentation — title, description, or the SERP itself |
| Steady | CTR steady | Down | The page — a genuine UX problem |
| Up | CTR down | Conversions down | Query mix — you are ranking for broader terms |
| Down | CTR down | Down | Site-wide — check availability and deploys first |
The fourth row catches the case that fools people most often. Traffic rose, conversions fell, and everyone concludes the page got worse — when what changed is that the page started ranking for looser queries bringing visitors who were never going to convert. Nothing about the page is broken.
Impressions are the only one of the three that your page design cannot affect. If impressions moved, the cause is upstream of the page, and no amount of redesign will address it. That single check settles most of these arguments in five minutes.
Why does the same chart fit both explanations?
Because sessions and conversions sit at the end of the chain and reflect everything that happened before them. A number at the end of a pipeline cannot tell you which section of the pipe is blocked.
Most reports show exactly that end-of-chain number and nothing else, which is why the meeting turns into a debate about interpretation. Adding impressions and CTR next to it converts an argument into a lookup, because those two are measured before the visitor ever reaches your page experience.
What does a presentation failure look like up close?
Position holding steady while CTR falls. Your result is in the same place and fewer people are choosing it, which means something changed about the surroundings or about your own snippet.
Three causes account for most of it. A competitor rewrote their title to answer the query more directly. Google added a feature above the results — a snippet, a pack, an AI-generated answer — that satisfies the query before your listing is reached. Or your own title and description were rewritten in a release and nobody connected the two events. All three are visible by searching the query yourself and looking at what sits above you, which is an underrated diagnostic step.
Comparing this month to last month and treating the difference as a finding. Seasonality, weekday mix and reporting-window effects all move these numbers. Compare to the same period a year earlier as well, and treat only a change visible in both comparisons as real.
When is it genuinely a UX problem?
When impressions and CTR are flat and the behavior after arrival changed. That combination has no other explanation — the same people arrived from the same queries, and something on the page stopped working for them.
At that point the useful next step is narrowing by device and template rather than opening a redesign. Mobile-only regressions usually trace to a layout or interaction change; a regression on one template while others hold points at whatever shipped to that template. Both are far cheaper to find than to argue about, and both produce a specific fix rather than a project.
What if all three are flat and leads still fell?
Then the problem is downstream of the site, and the site is the wrong place to look. This is the case nobody checks and it is more common than it should be.
Forms that silently stopped delivering, a phone number changed on the site but not in the CRM, a spam filter quietly eating submissions, a sales team that stopped logging inbound calls — each produces a fall in recorded leads with no change in any web metric. Submit your own form and watch the whole path before accepting that the page is at fault. It costs ten minutes and it occasionally saves a quarter of redesign work.
There is a reputational version of this too. A business whose E-E-A-T signals weakened — reviews turning negative, a key author page removed, credentials dropped in a redesign — can hold its rankings and its clicks while converting fewer of them, because visitors are arriving and deciding against you for reasons the analytics never records.
Before the next redesign is approved, pull impressions, CTR and one on-page behavior metric for the affected pages and put them in one row. Whichever number moved is the layer that failed, and the team that owns it is the team that should be doing the work.
Traffic down and nobody agrees why?
We will isolate which layer actually failed — rankings, presentation or the page itself — before anyone rebuilds anything.
Get in Touch →Does this apply to your site?
Reading about it is one thing. Point the scan at your own site and see whether this applies to you, and what it is worth fixing.
Free and unlimited. No account, no card, and you get every finding rather than a teaser.