There's a particular kind of frustration that comes from a search that works fine on paper. The index is populated, the engine is running, and results are appearing. And yet visitors aren't clicking, conversions are flat, and nobody can quite explain why.
The answer is almost always relevancy. Not whether results exist, but whether they're the right results for the person who asked.
What relevancy actually means
A relevant result is both the right answer and the obvious answer. From your side, you know it's correct. From the visitor's side, it looks like you read their mind.
That balance matters. You can have a result that's technically accurate but phrased so badly that nobody trusts it. You can have a result that sounds plausible but answers a different question entirely. Relevancy lives in the gap between what you know the answer to be and what the visitor is able to recognise as useful.
At its most basic, you measure relevancy through click-through rate (CTR). For e-commerce specifically, a combination of CTR and conversion rate (CVR) gives you a fuller picture, because a click that doesn't convert suggests the result looked right but wasn't.
What makes a visitor click
Consider a result with a 97% click-through rate and an average click position of one. What makes it perform that well?
The query is matched word for word in the title. The meta description beneath it is concise, accurate, and describes exactly what the visitor will find. There's no ambiguity, no mismatch, no gap between expectation and reality.
The number of results shown isn't the deciding factor. A single result that matches the query but has a misleading title or a vague description will still get ignored. Two results with perfect relevancy signals will both get clicked. The content of the result matters far more than the quantity.
The field of dreams problem
There's a tendency in search design to believe that building the result is enough. Curate the perfect product, write the ideal description, pin it to the top of the results page, and visitors will find it.
They won't. Not automatically, and not just because it's there.
Building a curated result is no guarantee that anyone will click on it. The result still has to communicate its relevance to the person looking at it. If the title doesn't reflect the query, if the description is too generic or too promotional, if there's any signal that the result is what you wanted to show rather than what they wanted to find, the click won't come.
Three patterns that kill click-through rates
False promises. Promoting items that bear no relation to what the visitor searched for. Even if the product is good, the mismatch signals that the system isn't listening.
Undisclosed advertising. Regulators forced Google to label promoted results as advertising for a reason. Visitors are quick to sense that a result is serving the site's interests rather than their own, and they skip it.
Over-curation that shows its hand. When results feel too controlled, too perfectly arranged, too obviously steered, visitors sense the manipulation. If the title and description don't reflect the actual search query, click-through rates drop sharply.
Relevancy as a design problem
None of this is accidental or random. Relevancy is the outcome of deliberate design choices: how metadata is structured, how titles and descriptions are written, how the search engine weighs different signals, and how results are presented.
Poor relevancy is rarely a technical failure. The engine is usually doing exactly what it's been told. The problem is that nobody checked whether what it was told matches what visitors actually need.
That's a fixable problem. It starts with understanding what your visitors are actually searching for, checking whether your results reflect that accurately, and adjusting the signals your engine uses to rank and display them.
When relevancy is right, visitors don't notice the search. They find what they came for. That's the goal.
If your results are showing but nobody's clicking
Start by looking at CTR data alongside the queries that generated each result. Where the click-through rate is low, ask whether the title and description accurately reflect what the visitor was searching for. Look for mismatches between the query language and the language used in the result.
Then look at whether any of the three patterns above are present: false promises, undisclosed promotion, or over-curation that's visible to the visitor. Each one erodes trust, and trust is what drives clicks.
Relevancy isn't a feature you configure once. It's a signal you read continuously, because what visitors need shifts over time, and your results need to shift with it.
Murmuration helps retailers and digital teams understand and fix onsite search. Get in touch if you'd like to talk through what a diagnostic might look like for your site.
▶ Know someone who’d love this? Forward it their way.
▶ Was this email forwarded to you?