Case Study. D2C E-Commerce
The Blog With Fourteen Million Impressions and Almost No Buyers
Quick answer
A medical apparel brand was generating close to fifteen million monthly search impressions and converting almost none of it. The blog was ranking for clinical and anatomical queries that shared vocabulary with the products but nothing else: the audience was students and patients researching conditions, not clinicians buying uniforms. The fix was to stop chasing that demand, and to build landing pages matched to the documented dress codes of individual large employers, a query type with near-zero competition and unambiguous purchase intent.
This is one of the clearest examples I have worked on of a metric that looks like success and is not. The brand sells clinical apparel direct to consumers. Its search dashboard showed nearly fifteen million impressions a month, a number most e-commerce operators would be delighted with, and a business that could not work out why organic revenue was so far behind the traffic story.
The answer was visible in one column of the report, and nobody had looked at it. The click-through rate on the blog was three tenths of one percent.
What the Impressions Actually Were
Pulling the highest-impression non-branded queries made the problem obvious in about ten minutes. The blog’s best-performing pages by impressions were ranking for clinical and anatomical terms: placental positions, chromosome pairs, cell structures, fracture types. Tens of thousands of impressions each, month after month.
Every one of those is a legitimate query with real volume. None of them is a person who buys clinical apparel. They are medical students revising, and patients looking up something a doctor said to them. The content was genuinely good and genuinely relevant to the query, which is exactly why it ranked. It was simply serving an audience that would never purchase.
| Query type | Monthly impressions | Commercial intent |
|---|---|---|
| Anatomical and clinical terms | Tens of thousands each | None |
| Condition and diagnosis lookups | Tens of thousands each | None |
| Product and category terms | A fraction of the above | High |
| Employer dress code terms | Very low at the time | Very high |
“Fifteen million impressions at a third of a percent click-through is not a traffic asset. It is a very large audience being shown something they did not want.”
Ram Kr Shukla, SEO and Growth Consultant
This is worth dwelling on because the diagnosis runs against instinct. The reflex when impressions are high and revenue is low is to improve conversion: rewrite titles, add calls to action, redesign the blog template. All of that would have failed, because the visitors were not underconverting. They were the wrong people.
The Uncomfortable Recommendation
I recommended deliberately losing most of those impressions. Stop producing clinical education content, stop optimising the pages that attracted non-buyers, and accept that the headline traffic number would fall, in exchange for a search footprint that reflected the actual business.
That is a difficult conversation, because impressions are the metric everyone above the SEO team is watching. A chart that goes down is a chart somebody has to explain. The argument that made it possible was reframing the measure: not how many people saw us, but how many of the people who saw us wanted what we sell.
The general version of this problem, and why it is so common, is in reporting on revenue rather than vanity metrics and blog traffic versus category traffic.
The Tactic That Replaced It
The replacement demand came from an observation about how this category actually works. Large employers in this sector have documented dress codes. Staff at one institution wear a specific colour, staff at another wear a different one, and new joiners are told to buy something compliant, often at short notice.
So they search for it. Not for the product category, and not for the colour, but for the requirement: what am I supposed to wear at this specific place, and where do I buy it. That is a query with total purchase intent and essentially no competing content, because building a page for a single employer feels too narrow to be worth anyone’s time.
Roughly a dozen of these went live. Within weeks they were ranking at average positions between one point five and six point three. On a competitive category term that would be a year’s work. Here it happened almost immediately, because the query had no incumbent.
| Signal | Before | After | Change |
|---|---|---|---|
| Total clicks | 75,217 | 81,394 | +8.2% |
| Branded clicks | 16,527 | 19,237 | +16.4% |
| Non-branded clicks | 58,690 | 62,157 | +5.9% |
| Site-wide CTR | 0.50% | 0.60% | +20% |
| Blog CTR | 0.30% | 0.40% | +33% |
| Total impressions | 14.8M | 12.8M | -13.7% |
| Collection page clicks | 13,169 | 14,093 | +7.0% |
| Home page clicks | 11,345 | 13,750 | +21.2% |
Read that table the way most dashboards would and it contains a failure: impressions down nearly fourteen percent. Read it correctly and it is exactly the intended result. Clicks rose while impressions fell, which is only possible if the impressions being lost were ones nobody was clicking. Click-through rate rose a fifth site-wide and a third on the blog. The traffic became smaller and considerably more valuable.
What I Would Do Differently
Two things. First, I would have set the reporting frame before touching anything. Because the impression decline arrived before the full benefit of the new pages, there was a period where the chart looked bad and the explanation sounded like an excuse. Agreeing in advance that impressions would fall, and that this was the plan, would have made that month far less uncomfortable.
Second, I underestimated how quickly the institution pages would rank, and built a dozen when the same research supported several times that number. The constraint was never demand or competition, it was how many employers had been researched properly. If I ran this again I would put far more effort into the mapping phase before building anything.
What Transfers
Two things transfer, and neither is specific to apparel.
The transferable lessons:
- High impressions with a very low click-through rate is a diagnosis, not an achievement
- Topic overlap is not intent overlap: sharing vocabulary with your product is not the same as wanting it
- Deliberately shedding non-converting visibility is a legitimate strategy, and needs agreeing before the chart moves
- Where a named institution has a documented requirement, a page answering that requirement faces almost no competition
- The research is the moat: anyone can copy the page template, very few will do the underlying mapping
- Answer the question first, sell second, particularly when the visitor arrived confused rather than ready
The institution page idea in particular generalises further than it looks. Anywhere a large organisation publishes or enforces a standard, and individuals have to comply with it at their own expense, the same search behaviour exists and the same gap is usually sitting open.
Related: e-commerce SEO, Shopify SEO, healthcare SEO, entity-specific landing pages, keyword research built around buyers, and the other case studies.
Common Questions
How can a blog get millions of impressions and produce almost no revenue?
By ranking for queries that share vocabulary with the product but not intent. A medical apparel brand publishing clinical explainers will attract students and patients researching anatomy, not clinicians buying uniforms. Search engines match the topic correctly; the traffic is simply the wrong audience, and the click-through rate tells you so long before the revenue does.
Is a falling impression count always bad?
No, and treating it that way causes real damage. If impressions fall while clicks hold and click-through rate rises, you are losing visibility on queries that were never converting. That is the index correcting itself. The number to watch is not impressions, it is clicks and revenue per thousand impressions.
What are institution-specific landing pages?
Pages built for the exact requirement of a named employer rather than for a generic product category. In uniform retail, large employers have defined dress codes, so a page answering what a particular institution’s staff are required to wear, and showing only compliant products, matches a search that a category page cannot.
Why do those pages rank so easily?
Because almost nobody builds them. The queries are specific, low in volume individually, and commercially unambiguous, which means very little competing content exists. Pages of this type reached the top of page one within weeks, which is close to unheard of on a competitive category term.
Does this tactic work outside uniform retail?
Anywhere a named institution has a documented requirement and buyers search for it. Airline and hotel staff outfitting, school uniform suppliers, security and hospitality workwear, industrial safety equipment with site-specific standards, and parts fitted to a particular machine model all share the shape.
What was the actual outcome?
Within the reporting period, clicks recovered while impressions fell sharply, click-through rate rose by a fifth overall and by a third on the blog, and the newly built institution pages entered at average positions between one and six. The impression decline was the intended consequence of pruning demand that never converted.
Is your traffic number hiding an intent problem?
High impressions with a low click-through rate usually means you are ranking for the right topic and the wrong audience. That is diagnosable in an afternoon and it changes what you should be publishing.
Tags: Case StudyE-Commerce SEOSearch IntentShopifyLanding Pages
