How strong is this evidence?

What is a benchmark, and what is it not?

A benchmark here is a published range for an operating figure, describing what is normal across many coaching practices. It is a diagnostic instrument: the question it answers is whether your own number is inside or outside the range, and nothing more.

It is not a target. A close rate at the top of its band is not automatically better, because the bands are not quality measures. A 70% close rate on warm referrals usually means a small number of extremely warm leads rather than an excellent sales process.

It is not a forecast either. Nothing on this site predicts what your business will do. Benchmarks describe a population, and any individual practice can sit outside a band for reasons the population average cannot see.

Which figures are measured, and which are estimated?

Every headline figure on the site, its source, and what kind of evidence that source is. Read the right-hand column before quoting anything from the left.

Sixteen headline figures: five measured, ten practitioner estimates, one mixed. Of the five measured, three come from a profession-wide census and two from a cross-industry email study. The estimate rows carry the most-read pages on the site, which is precisely why the labels stay on.

Why is the acquisition data so much weaker than the rest?

Why publish estimates at all?

Because the alternative is not better information, it is the same information with the labels removed. Coaches are already making decisions against these numbers; they are circulating in courses, forums and sales pages, usually stripped of any indication of where they came from.

A labelled estimate is more useful than an unlabelled one and considerably more useful than silence. It lets a reader apply the right amount of confidence, which is the entire purpose of publishing a band rather than a point.

The line we hold is that an estimate is never presented as a measurement. Anything derived by us from other figures says so in the caption. Anything with no published rate at all is left empty rather than filled, which is why three of the eleven funnel structures carry a documented shape and no conversion number.

How is a band constructed?

Four rules, applied to every figure on the site. They are unexciting deliberately, because the aim is that two people following them would produce the same band.

1. Take the range as published, not an average of it

   Where a source gives 40% to 70%, the band is 40% to 70%. Collapsing it to 55% would invent a precision the source does not have, and a midpoint is wrong for almost every reader.

2. Never merge sources of different strength into one figure

   A measured figure and an estimated one are never averaged together. Where both bear on a topic, the page reports each separately and says which is which, as on the retention page.

3. Label every derivation in the caption

   Several useful figures on this site are arithmetic rather than sources: hours per client, effective hourly rate, per-subscriber revenue, the two-year runway cost. Each one names the figures it was derived from in the caption under the table.

4. Leave the column empty when there is no figure

   Three funnel structures have a documented shape and no published conversion rate. Those pages say so. A plausible number in an empty column is the most damaging thing a reference site can publish, because it is indistinguishable from a real one.

What is a single documented case actually worth?

The Google Ads figure on this site traces to one case: $2,800 of spend, 43 booked calls at $65 each, seven clients, $17,500 of first-month revenue. A 6.25× return, and genuinely encouraging.

It is also one case, with a defined niche, a landing page converting at 17%, and a strong call process. All three had to work simultaneously, and the case does not tell us how many coaches ran similar campaigns and did not publish the results.

Published case studies are selected for being good. That is survivorship rather than dishonesty, and the correct response is not to discard the case but to read it as proof that the channel can work rather than as the rate you should expect. Close rates on paid search sit at 14% to 18% generally, well below the referral band.

What conflicts of interest does this site have?

This site is published by Coachful, which makes coaching software. Coachful appears in rankings here marked as ours, and it is the recommended pick where a page recommends software. That is a direct commercial interest in some of what is published, and it is disclosed on the page rather than only in the footer.

Several pages carry that interest more heavily than others. The tool stack costs page argues that fragmented stacks are expensive. The time page argues that administrative hours suppress the effective hourly rate.

Both conclusions favour our product, and both rest on the same 6 to 10 hour estimate from the report rather than on data we generated.

Three commitments limit the damage. No affiliate links anywhere, including to our own product. No paid placement, ever, from any company.

Every competitor named gets at least one thing it genuinely does better named with it, because ranking our own product first is a judgement a reader can weigh and misdescribing somebody else’s product is a lie.

What would change our mind: a measured study of coaching acquisition cost, a coaching-specific email benchmark with a comparable sample to GetResponse’s, or a measurement of administrative hours by tooling type. Any of the three would replace an estimate on this site with a figure, and we would rather publish the figure even where it contradicts us.

How should you cite this site?

Quote anything here without asking. Four requests, all of which are about the number surviving the journey intact.

• Carry the estimate label. A practitioner guess that arrives in someone else’s article stripped of its uncertainty is how soft figures become industry facts.
• Quote the band, not a midpoint. "40% to 70%" is the finding. "55%" is a number nobody published.
• Name the original source where one exists. ICF Global Coaching Study 2025 and GetResponse 2024 did the work; we transcribed it.
• Keep the qualifiers. "One to three months" without "with active outreach" is a different and much less true claim.
• Send corrections. Measured beats estimated every time, and we will say where a replacement figure came from.