The conversion statistics in this category do not survive a check

If anyone has pitched you a quiz funnel in the last year, you were shown a multiple. Two times the conversion, four times the leads, or one particular figure that keeps turning up with a decimal point in it. We were going to put one on this site too. Then we went looking for the study underneath it, and the study is not there.

This post is the working. It is also why you will not find a conversion-lift number anywhere else on this site, including on the pages where having one would help us sell.

What a usable statistic would need

Three things, and all three at once. A publisher with no commercial interest in the answer. A stated method, so you can see what was actually measured. A denominator, so the multiple has a base. Nothing in circulation had all three. Most of it had none, which makes it useless for deciding where a pound of spend goes.

The three claims everyone is quoting

The one with the decimal place

The most quoted figure in interactive-content marketing is a lift multiple written to one decimal place, usually as 16.9x. It appears in vendor blogs, agency decks and conference slides, almost always with no link. Follow it back and it attributes to a research organisation that leaves no independent trace of existing. No site, no other published work, no registration. We are not saying the number is wrong. We are saying there is nothing underneath it to be right or wrong about.

The 2016 study

The second is a real study from 2016, sponsored by a vendor selling into the category. Read its method and the problem arrives immediately. It surveyed marketers on whether they believed interactive content performed better than static content. It measured opinion. It is quoted as though it measured performance.

A room full of people who bought a thing saying the thing works is a customer-satisfaction survey. There is nothing wrong with one until somebody relabels it as evidence.

The completion rates

The third group are completion figures: what share of people who start an assessment reach the end of it. These are the most defensible numbers available in this category, and they are still not independent. Every one we could trace was published by a company that sells the software, measuring its own accounts, with no control and no disclosure of which customers were left out. A vendor's own aggregate tells you something real about its best customers. It tells you nothing about your traffic.

Our own research hit the same wall

Our market study went at the same question from the other side and failed, in a way worth reporting. Across a September 2026 sweep of 13 competing platforms and a census of 174 published templates, it could not find one independent cross-format conversion dataset. Not a single comparison of one instrument against another against a static page, run by anyone with nothing to sell.

Three more things the same study set out to establish and could not: installed-base counts for any platform in the category, deal value and lifetime value by B2B vertical, and the partner-network size of the largest vendor, whose own two published user figures contradict each other. The research page publishes those failures in full, because a study that reports only what it proved is an advert with a chart on it.

Why it keeps getting quoted

Because the honest alternative is harder to sell. A lift multiple is a promise that fits on a slide, goes straight into somebody's business case, and turns into a spend decision nobody revisits. What we have instead is an argument about when the buying decision actually happens, and an argument takes a paragraph.

And because almost nobody checks. The figure has been in circulation long enough to cite itself. A vendor quotes an agency, the agency quotes a blog post, the blog post quotes a deck, and the deck has the decimal place that makes it look measured.

What we use instead

Four statistics, all from publishers with no stake in interactive content, all with a sample size we can state out loud. None of them says a diagnostic converts better than a landing page. They say something more useful about where your pipeline actually comes from: by the time a buyer fills in anything at all, most of the decision has already happened without you.

Two more sit behind those. Buying groups average 13 internal and 9 external participants (Forrester, 2026), and over 40% of deals stall on internal misalignment rather than on price or product (Edelman/LinkedIn, 2025). Together they describe a committee that does most of its work before your sales team hears a name, and a stall that ties up pipeline for a quarter without closing anything.

Those four describe the market. They do not describe our results, and no sentence on this site will let them do that by sitting next to a claim of ours. What we make of that window is our own argument, carried without a borrowed number, and you can judge it on whether it holds.

Three questions for the next deck that shows you a multiple

1. Who paid for it? A vendor-funded study is not disqualified. It is discounted, and the discount is large.
2. Opinion or behaviour? “Marketers report” and “visitors did” are different sentences of roughly the same length. Only one of them is a measurement.
3. What is the denominator? Four times what, against which page, over what period, across how many visitors. A multiple with no base is a mood.

If a deck gets through all three, send it over. We would genuinely like to read it, and we will publish it here with the correction.

The only conversion number that is yours

One figure in this whole argument is worth having, and it is the one your own traffic produces. You already hold the inputs. Last month's spend, the sessions it bought, and how many of those your sales team qualified rather than filed. That is a real denominator, and it is the only one with your business in it.

So the honest version of this pitch reads like this. We do not know what a first build will do to your pipeline, nobody credible does, and running it is how you find out. What we can do is agree what gets measured before it goes live rather than after, which is the part most agencies leave until the number arrives and can be framed.