How we use numbers
Four kinds of number matter here, and we label which is which. Three of them appear on this site. The one that does not is the only one we would actually base a recommendation on, and it does not exist until we have measured your business.
Independent research — a named study, a named sample, a named date. We say who ran it and when, so you can check it.
Vendor data — a number produced by a company that sells the fix. Sometimes it is the best data that exists, because they have millions of records and no researcher does. We use it, and we label it.
Your measurement — what we timed, at your business, with the date on it. This is the only kind we would ever base a recommendation on.
And one kind that is none of the above: a number we invented, for an illustration. Where we do that, the illustration says so and so does the picture it sits in. None of it is a benchmark for anything.
Numbers we refuse to use
We traced these and they did not survive. They are all in wide circulation, and if you have heard one from somebody selling you something, you now know as much as they do.
The widely quoted figure for unanswered small-business calls comes from a marketing company in 2016.
"Most clients hire whoever answers first" has no published method behind it anywhere. Every citation chains to another blog.
The annual-loss totals for missed calls have no locatable source at all.
"95% of AI pilots fail" comes from a 2025 working paper that was not peer reviewed, and the figure referred to projects showing no measurable return rather than projects that failed.
"80% of knowledge bases are out of date" traces to a vendor blog with no study behind it.
"30% to 50% of consulting recommendations are never implemented" has no study anywhere that we can find.
Per-missed-appointment dollar figures, which pile three assumptions on top of each other and present the result as a finding.
One large dataset on patient calls, which we would use if its original publisher can be named and linked. Until then it stays off the site, even though it is the best-sampled thing available, because we cannot ask you to check something we cannot cite.
Where our own numbers come from
When we publish something about a study, the source and the date are beside it. When we publish somebody's market price, it is inside a block labeled as market rates, with the source and the date.
We publish no prices of our own, and that is deliberate. We name a fixed quote after we have seen the work, because a price published before we have seen anything is a guess with a number on it.
If you find something on this site that does not meet that standard, tell us and we will fix it or take it down.
Hold us to this
The paragraph above is not a formality. If a number anywhere on this site does not meet the standard on this page, tell us which one. It gets a source or it comes down, and we will say which.