Guide · Disclosure
How to Publish Funnel Numbers Honestly
A funnel number is only useful if a reader can tell where it came from. Every published rate should state its definition, sample, period, market, source system and rounding, and every modeled number should say it is a model.
Step 1: Attach six facts to every rate
| Fact | Example |
|---|---|
| Definition | Registrations per landing visit, step rate |
| Sample | Counts behind the rate, such as 1,224 of 6,800 |
| Period | October 1 to 31, 2026, closed November 4 |
| Market and product | Great Britain, online casino |
| Source system | Analytics for visits, server for registrations |
| Rounding | Counts rounded at each stage; rates to one decimal |
The example counts in the table are invented to show the format.
Step 2: Do not call one source a benchmark
A rate from one site, one market or one period is a data point, not an industry benchmark. Calling it a benchmark invites readers to judge their own sites against it. If you publish a range from several sources, show how many sources, how they were measured and whether the definitions matched. If you cannot show that, publish the individual figures with their facts instead of a range.
Step 3: Label models plainly
Models are useful for teaching and planning. Label them where the number appears, not only in a footnote. This site uses the words "illustrative model, not measured data" next to every modeled figure, including the funnel explorer outputs. A chart without that label invites someone to screenshot it as evidence.
Step 4: Avoid claims a reader cannot check
Google's guidance on helpful content asks whether content provides "original information, reporting, research, or analysis" and warns against "fabricating creator profiles." For funnel content, the same idea applies to numbers: a published figure should either come with its sources and facts or be clearly marked as an example. Phrases such as "industry average" or "most operators see" need evidence behind them, or they should not appear.
Step 5: Respect what the data may reveal
Funnel counts can identify small groups, especially for one market and a short period. Aggregate before publishing, remove counts below a threshold you set in writing, and never publish anything that could identify a customer. Check commercial agreements before naming any partner's figures.
Step 6: Keep a correction route
Publish how readers can report an error, and record corrections with dates on the page. This site's editorial policy explains how its own corrections work.
Checklist
- Six facts attached to every rate.
- No single-source figure called a benchmark.
- Models labeled where the number appears.
- No unchecked "average" claims.
- Small counts suppressed; no customer can be identified.
- A correction route published.
Questions
Can we publish our own funnel rates as an example for others?
Yes, with the six facts attached and with permission from anyone whose data is included. Call them your site's figures, not a benchmark.
Is it acceptable to round rates for readability?
Yes, if you state the rounding and keep the underlying counts available. Readers rebuilding the figure need to know the rule.
Why does this site not publish benchmark ranges?
Because it has no operator data measured the same way across markets. The scenario library offers editable assumptions instead, labeled as such.
Sources
Each source was read on October 27, 2026. How we choose sources is set out on the sources and reading list page.
- Google Search Central, Creating helpful, reliable, people-first content, last updated October 5, 2026.