A number is usually unpublished because it is expensive to collect
468 words · about 2 min
If a number is not published anywhere, that is usually because gathering it costs a lot — so "nobody has it" and "we could get it cheaply" pull against each other. Testing how many sources publish the number, before doing anything else, kills a bad idea in two hours instead of two weeks.
- How it was measured
- Worked out from four candidate projects that passed every other check and then failed on the cost of gathering the data. That test now runs first, against thresholds we write down before we look.
- Last checked
Listen to this Playing read from this page, nothing added
Watch it plot 33s · narrated · the same figures as above
The appeal of a number nobody publishes is obvious: nobody has it, so putting it together makes something new. The trap is just as obvious once said out loud, and almost never said.
A number is usually unpublished BECAUSE it is expensive to collect.
"Nobody has published it" and "we can get it cheaply" pull against each other, and any project worth doing lives in the narrow band where both somehow hold. Most candidates that look like they are in that band are not — they are simply expensive, and the expense is what has kept everyone else out.
The cheapest way to kill an idea
Take forty sources and count how many publish the number without you having to phone anyone.
| How many publish it | Verdict |
|---|---|
| 50% or more — half or better | You can build the series. The only barrier is the tedium. |
| 25-50% | Marginal. Work out honestly what gathering it will cost before committing. |
| Under 25% — fewer than one in four | Dead, and dead for the right reason. |
Two hours and no money, against a full survey of who else is in the market. Run it before anything else, because a candidate that fails here fails no matter how good the rest of it looks.
Check how fast the number goes stale, too
A second cheap way to kill an idea: how often does the number actually change?
Two candidates died on this alone. One moved about 6% a year, another changed once a year — which means a six-month-old figure is inside the range you would get by chance anyway. A number that changes once a year cannot support a product people come back to. You update it once and it sits there for eleven months, and nobody returns.
What you want is a number that decays steadily, where being three months out of date is visibly worse than being current. That is what makes somebody check again.
And when an idea does clear both, size is not what makes it believable
The best-known published figure in one market we looked at rests on 282 replies, a 7% reply rate, three years of age, and 93 downloads. A competitor advertising a weekly update had left 98.6% of its pages untouched for four months, with its flagship page showing the words "may be outdated".
Nobody is checking any of it. Which means the bar is not how much ground you cover:
"44 of 44 businesses in this named market, checked on this date" is more defensible than "23,000 properties", and it is what nobody currently offers.
Every single one in a named market, with the date you checked. It is cheaper than scale and worth more.
Tracked as OPACITY-PARADOX, CREDIBILITY-IS-CENSUS. If a number here looks wrong, tell us and we will check it against the source.