Thirteen formats, and what picking the wrong one costs

Every interactive asset in B2B is one of 13 things. We arrived at that number the slow way, classifying 174 published templates across 23 fields for our own market study, and the list held all the way through. Prewritten personality tests at one end, peer benchmarks at the other, and eleven in between. Which one you pick decides how much of next year's spend you can defend.

Which one to build usually gets decided on how engaging it sounds, or on what the marketing team fancies making. That is the wrong axis. The one that decides whether you still own an asset in two years, and whether it is still producing qualified leads, is what the thing runs on.

Authored or fed

There are two kinds. Authored, where the output was written in advance: an authored type test picks one of five prewritten results, a checklist is a list. The value sits in the writing, and writing is now free.

Fed, where the output is computed from something you hold and a competitor does not. A grader crawls the real asset. A benchmark places the respondent against a cohort of real answers. A pricing calculator runs their inputs through your actual rate card.

AI collapsed authoring and has not laid a finger on data. A cheap builder will write eleven decent questions, five result pages and a scoring rubric before lunch. It cannot crawl your prospect's estate, and it cannot conjure four hundred peer responses that do not exist. Which is why two of the 13 will never again bring you a lead you could not have had anyway.

The thirteen, and what each one runs on

Three that hold, two that do not

Three run on data you own and cannot be reproduced by writing harder: the grader, the benchmark tool and the survey. Our research scored those three lowest for exposure to AI commoditisation, and it was not close. They are also the three where the leads keep arriving after the launch quarter ends.

Two run on nothing: the self-assessment checklist and the prewritten personality test. Both are cheap, both work as a first step, and both are now a list and a lookup table with a good design on them. Build one if you want traffic this month. Do not build a practice on one, and do not put real spend behind it.

The scorecard, which is what most people mean when they say quiz funnel, sits in the middle and deserves its popularity. It is the only one in the taxonomy that generates leads, qualifies them and segments them at the same time. It is also thin the week a competitor publishes their own, unless there are genuine benchmarks underneath the bands. Thin means the same traffic, fewer qualified leads, and a sales team back to guessing.

What the market has actually built

This is the row of counts that should decide your roadmap. Across 174 templates in the largest public library in the category, there is one calculator, one benchmark tool, one maturity assessment and no graders at all. Against 34 self-assessments and 26 scorecards.

Supply is piled into the two things a cheap builder now reproduces, and absent from the three it cannot touch. Which means most of the spend in this category is being pushed at the copyable half. The same sweep looked at 13 competing vendors and found that not one of them offers peer-cohort benchmarking. “How do we compare to firms like ours” is the highest-value question a B2B diagnostic can answer, and nobody in the category answers it.

The pair that compounds

One move in the taxonomy is worth more than the rest, and it is a pair rather than a build. Run the survey first: it manufactures the dataset. Then the benchmark rents that dataset back to individuals, indefinitely, and every new response makes the cohorts denser.

Year one you publish a research report, earn the links and collect the leads. Year two the same data is a live comparison nobody can copy without repeating year one, and it keeps returning pipeline while your competitors start from nothing. Template libraries ship both halves and never connect them, which is the most expensive omission in the category.

What picking wrong actually costs

Not the build. The build is the cheap part and costs roughly the same whichever of the 13 you choose. The cost is everything you push at it afterwards: the spend, the links, the sales time, a quarter of positioning, and the rewrite when a competitor ships the same thing in March. That is a year of traffic buying leads nobody can close.

The opposite mistake is just as expensive. A benchmark with no dataset behind it is a cohort comparison against 11 responses, and showing someone their percentile against 11 strangers damages trust faster than offering no comparison at all. If the data is not there yet, the honest build is the one that goes and gets it, even though it returns fewer leads in the first quarter.

Four things we would settle first

1. Do you already hold data nobody else does? A rate card, a crawlable asset, an owned list large enough to survey. If yes, build on that and stop reading lists like this one.
2. If not, can you go and make some? A survey is the cheapest way to manufacture a dataset. Budget for it as the asset itself, not as a campaign that runs once.
3. Is your ICP written down? If it is not, a qualification quiz is not available to you yet, whatever the software says. The routing rules have to come from somewhere.
4. What decays? A readiness assessment on a real deadline brings the highest intent per completion in the set and has the shortest shelf life in it. That is fine, as long as everyone knows they are funding 12 months of pipeline rather than an asset.

Those are the four we would open with, not a rule. The right answer depends on what you already hold, and the fastest way to find out is to say out loud what data your business generates that your competitors never see.

If none of it is obvious, that is worth knowing too. It usually means the first build is the survey, the compounding asset is the second one, and the leads from the first pay for the second.