Your best-performing page is probably not your best page.

Sort your landing pages by sessions and look at the one on top. In a lot of service businesses it is a definition post, a cost guide, or a how-it-works explainer that pulls thousands of visits and produces almost nothing. Then look further down, past the fold of the report, where a service page gets a fraction of the traffic and most of the leads.

Nothing is broken in that picture. It only becomes a problem when the traffic number is the one being reported, celebrated, and used to decide what to write next.

Four kinds of queries, four different jobs

Search intent gets taught as a four-box model and then applied as though only one box matters. All four have a job. They just pay out on different timelines.

Problem unaware. Someone searching why their commercial HVAC unit short cycles does not know yet that this is a service call. They are diagnosing. This content builds familiarity and earns citations, and it converts almost nobody today.

Research. How often a rooftop unit should be serviced, what preventive maintenance covers. The person now knows a vendor exists. They are learning what normal looks like so they can tell whether they are being overcharged later.

Comparison. Contract versus per-call service, in-house versus outsourced, one vendor against another. This is where a decision is being formed and where most businesses have the least content, because comparison posts feel uncomfortable to write.

Ready. Commercial HVAC service plus a city name. Emergency repair near me. Pricing. Small volume, high value, and almost always the pages with the fewest words on them.

The error is treating rankings on the first tier as interchangeable with rankings on the fourth, because both appear in the same report under the same heading.

A worked example

Say your top pages look roughly like this over ninety days.

Total sessions, 7,930. Total leads, 95. Site conversion rate, 1.2%, which looks unhealthy and is not.

Read the rows instead of the total. The two top pages produce 86% of the traffic and 11% of the leads. The bottom three produce 14% of the traffic and 89% of the leads. The comparison post converts at 3.4% and the two ready-intent pages convert at 11% and 16%.

Now the obvious question. Which row deserves the next piece of work?

Almost every content calendar answers by writing more of row one, because row one is where the traffic is and traffic is what gets reported. The right answer is usually row three, because 640 sessions converting at 3.4% is a page that would produce meaningfully more leads with meaningfully more visibility, and because comparison queries are the thinnest content category on most service sites.

What to do with each tier

Ready-intent pages get the technical and link investment. Small traffic, enormous value per visit. These are the pages worth internal links from everything else on the site, and the ones where a page speed problem or a weak local signal costs the most.

Comparison pages get the writing effort. This is the highest-return category on most sites we audit and the emptiest. If you have never published an honest comparison of your own approach against the alternative, that is the gap.

Research pages get maintained, not expanded. They already work. Keep them accurate, link them downward to the comparison and ready pages, and stop adding to the pile.

Problem-unaware pages get left alone, and kept. More on that in a second, because the instinct at this point is to delete them and it is the wrong instinct.

The mechanism that makes this work is internal linking between tiers rather than volume within one tier, which is the actual point of organizing content into topic clusters rather than publishing more of whatever ranked last.

Do not kill your informational content

Here is where the traditional advice has gone out of date.

Informational content converts poorly and always has. It is also, now, the content that AI search engines actually read. Ahrefs found that 99.2% of keywords that trigger an AI Overview have informational intent, which means your problem-unaware and research pages are the ones deciding whether an engine knows your business exists and describes it correctly.

So those pages have quietly changed jobs. They were traffic generators. They are now the substrate AI engines summarize when someone asks a question in your category, which is a different kind of value and one that shows up nowhere in a conversion report. We covered the mechanics of that in what makes content citable by AI engines.

Keep them. Maintain them. Just stop measuring them against a lead target they were never going to hit, and stop writing the eleventh one while the comparison tier sits empty.

Change what the report shows

The fix here is mostly a reporting change, and it takes an hour.

Tag every landing page with its intent tier. Then report sessions and conversions by tier instead of by page. The month you do this, two things usually become obvious at once: that the traffic growth everyone has been celebrating came almost entirely from the tier that converts worst, and that the tier producing the leads has not been touched in a year.

That is a visibility problem on the pages that matter, not a conversion problem, and it needs search work aimed at commercial queries rather than another round of page redesigns. If the tiers look healthy and the leads still are not arriving, the leak is further down the path, and the full diagnostic from query to closed deal covers where else to look.