The SEO Screenshot Problem

The graph that goes up and to the right

If you spend any time on LinkedIn, you’ll have seen the graphs. A screenshot of a visibility or traffic chart, arrow pointing sharply upward, captioned with something about phenomenal growth or the results a client is seeing. They’re everywhere, and they’re persuasive, because a graph going up looks like proof.

The trouble is that a lot of these graphs aren’t showing what they appear to show. They’re not fabricated, they’re usually built from platforms that estimate performance rather than measure it, and those estimates can move a long way without a business seeing a single extra enquiry.

This isn’t a dig at any particular agency or tool. It’s an honest look at how SEO reporting actually works, where the numbers can mislead even well-intentioned people, and why we report differently at Crucible.

Where these graphs actually come from

Most of these graphs come from platforms like SEMrush, which track where a website ranks for thousands of keywords and turn that into a single visibility or estimated traffic score. It’s a genuinely useful metric, but it’s built on an assumption.

To convert a ranking position into an estimated number of clicks, these tools apply a standard click-through-rate curve: roughly, position one gets around a third of clicks, position five gets a few per cent, and so on down the page. That curve is applied the same way across every keyword a site ranks for, regardless of what that keyword actually is or how someone searching it behaves.

That’s a reasonable shortcut for a broad, high-volume term. It’s a much less reliable one for a long-tail, niche phrase that a handful of people search for each month, where real click-through behaviour looks nothing like the average curve the tool assumes.

Why niche keywords make the graph look better than it is

We saw this play out recently with a client in a specialist trades sector. As part of the SEO work, we picked up rankings for a wide spread of long-tail terms, specific treatment types bundled with property-related phrases that, individually, only a handful of people search for each month.

Because the estimation tool assumes a consistent click-through rate across every one of those terms, its visibility score rose sharply. The graph looked like the account had taken off. When we checked that against Google Search Console and our own tracked keyword clusters, the real picture was more modest: genuine ranking improvement and more impressions, but nothing like the dramatic curve the estimate suggested, and no equivalent jump in enquiries.

That’s not a criticism of the work itself, the rankings were real. It’s a reminder that an estimated visibility score and actual business impact are two different things, and a screenshot only ever shows the first.

Search Console isn’t automatically clean either

Google Search Console is closer to reality. It shows actual clicks and impressions, not an estimate. But taken as one big total, it isn’t a clean read on its own.

Take our own site as an example. Crucible used to work across a much broader mix of industries, including a number of health and science-sector web builds. We’ve since focused on construction, property and manufacturing. Some of our older, high-performing content from those earlier sectors is still ranking, and it still shows up in our overall Search Console totals.

None of that traffic is relevant to what we do today. Reporting total clicks and impressions without filtering it out would flatter us with results that have nothing to do with our current business, the same way a client’s account can pick up one-word or off-topic matches that happened to rank by accident. Unless you specifically filter down to the pages and terms that matter, the total figure can be quietly padded by noise like this.

What we track instead

This is why we don’t rely on a single headline number from any one platform. Alongside Search Console, we use SE Ranking to track specific, agreed keyword clusters built around what a client actually does, for example a separate cluster for each core service line, plus a trust-signals cluster we run for every client, covering brand and reputation-related searches.

That lets us report on movement within the terms that genuinely matter to that business, rather than an aggregate that includes irrelevant long-tail noise or legacy content. It also means that when we tell a client their visibility is improving, we can show exactly which terms and which service areas that improvement sits in, not just a line trending upward.

SEMrush isn’t the villain here

None of this means SEMrush, or tools like it, aren’t worth using, they are. Visibility estimates are a genuinely useful early signal, and we use them ourselves. The point isn’t to throw that data out, it’s to make sure we’re never looking at it in isolation. We look at Search Console, SE Ranking and visibility-estimate data together, rather than treating any single platform’s snapshot as the full story.

It’s also worth saying these tools can get it wrong in the other direction too. We recently worked on a brand-new website in a very niche, highly localised subject, the kind of business that only needs a handful of enquiries a year given what it sells. On SEMrush, it looked like there was next to no organic visibility at all. Search Console told a different story: real, meaningful growth from a standing start, exactly what we’d expect for a site and a market of that size. Visibility tools can underestimate a site for plenty of reasons, weak crawlability, page speed issues, or simply a new site that hasn’t built up the history these platforms rely on. Looked at in isolation, either extreme, over-optimistic or overly bleak, is misleading. Together, the full picture is far more reliable.

Remembering the actual goal

None of this makes an upward graph meaningless. Movement in the right direction is still a good sign. But a graph is a proxy, not the goal itself. The thing that actually matters is the right people finding a business online and going on to enquire.

That’s the standard we hold ourselves to. Getting someone to click through is only half the job, the rest is making sure that visit has a genuine chance of turning into a lead. An agency optimising for an impressive screenshot rather than for that outcome might keep a client for a renewal or two, but sooner or later the client notices that the graph never turned into more business.

We’d rather build the kind of results that keep clients for years, not just the quarter after a good-looking report.