A calculator page has room for 1-2 ads. A written page has room for 6-12.
655 words · about 3 min
A page whose whole job is one calculator has room for 1-2 ads inside the content, against 6-12 on a 2,000-word article. That is a ceiling built into the format, and it holds whatever the subject. We claim no revenue gap alongside it: the earnings figures we have are one site each, so they measure two businesses rather than two formats.
- How it was measured
- The ad counts come from looking at the pages themselves, which is the part anyone can repeat. The two earnings figures come from marketplace listings where both traffic and revenue were checked by the broker, but they are ONE site each — bmicalculatornhs.co.uk at $12.72 and stovemastery.com at $43.29 — so they measure two businesses, not two formats.
- Last checked
Listen to this Playing read from this page, nothing added
Watch it plot 37s · narrated · the same figures as above
A tool page has a genuine advantage over an article: it is automatically unlike anything else. It has its own inputs, its own workings, its own answers, and nobody can spin a copy of it out of somebody else's writing.
We chose the format on that basis. That is true, and it is the wrong question.
Money from ads is usually quoted per thousand pageviews, and the trade calls that figure RPM. A site on $15 RPM earns $15 for every thousand pages people load.
| RPM | Ads inside the content | |
|---|---|---|
| A page that is one calculator | $12.72 | 1-2 |
| A long written page | $39-43 | 6-12 |
A site of calculators has roughly a sixth of the space to sell, and the reason is the shape of the page rather than its quality: there is less page.
This page originally said the gap was 3x in revenue, and that part was wrong.
The two RPM figures above are one site each — a UK BMI calculator at $12.72 and a stove site at $43.29 — and this page presented them as though they described two formats. They describe two businesses.
Set against our own spread across 123 dual-verified businesses, the calculator figure is unremarkable: a quarter of sites earn less than $6.88 and half earn less than $15.21, and $12.72 sits between those. The written-page figure sits above the level only a tenth of sites beat ($35.05). So the comparison put an ordinary site next to an exceptional one and read the difference as a property of the format.
It is the same error this site published a correction about in the RPM note — quoting a figure from near the top of a range as if it were the middle — committed here in our own copy a week later. The ad counts below were counted directly and are unaffected.
Two things are going on, and only one is the obvious one
The first is what happens across a visit. Advertisers bid most on the first page somebody loads, so on a site where people read several pages, the money per visit runs 20-25% above the money per page. A page that answers a question and gets closed has about 1.0 pages per visit, and gives up that premium entirely.
That is a 0.8x factor on its own — a fifth off. The bigger thing is space. A 2,000-word article has room for six to twelve ads inside the text. A calculator has room for one or two, above the tool and below it. That ratio was counted rather than reasoned out, and it does not depend on which two sites we happened to have revenue figures for.
What we should have asked
The advantage we picked the format for was real. The question we never asked was what it cost — and while we have not established the size of that cost, the direction is not in doubt: fewer ad slots per page is fewer ad slots per page.
A tool is still worth building. It just belongs inside a written page rather than being the page. One calculator in two thousand words carries the ads; a bare calculator page does not.
A warning about the whole category, found alongside it
While checking who else was doing calculators, we found something worth saying separately: of the top ten results for one common calculator search, six domains were under sixteen months old — registered within weeks of each other, against sites from 2005, 2009 and 2013.
Two readings, both true. Getting in is trivially cheap, which means nothing built there can be defended. And it sits squarely in the path of Google's action against mass-produced content, which is a test of what you were trying to do rather than of how much you published.
Tracked as FORMAT-SETS-CEILING, CALC-RPM-VERIFIED. If a number here looks wrong, tell us and we will check it against the source.