A mechanism that gets senior professionals to write down what they need

The median response to a business-to-business survey is 8.18% (Survicate, 2025). That is the number our own figure should be read against, with one honest caveat stated up front: an intake completed at the point of registration is not the same instrument as a fielded survey, and the two are not like for like. The comparison is directional, not an equivalence.

The harder problem in this market is not the response rate anyway. It is incidence. Cost per completed response rises from roughly US$7 where seven in ten people screened qualify, to roughly US$50 for an executive where three in a hundred do (Greenbook, which carries no publication date, so read it alongside the dated Veridata Insights, 2026, which puts narrow business-to-business targeting at 2 to 10% incidence, or about twenty people screened for every one who qualifies). Every one of those nineteen screened-out people is paid for.

Then there is the incentive. Research with executives runs at US$200 to US$500 an hour (Great Question, 2026), which is the price of interrupting someone whose hour is worth defending.

Set against that, two costs on our side are close to zero, and for a structural reason rather than a clever one. Our incidence is effectively 100%, because the organisation whose members these people are has already done the screening. It spent years assembling a list defined by sector, seniority and geography, and it verifies that list every renewal. We are not looking for the one in twenty who qualifies. Everyone on the list qualifies, by construction.

And our incentive cost is nil, because the thing the respondent receives is something they wanted anyway. They are going to the event regardless. The answers turn a room of strangers into a short list of named people worth finding, which is the outcome they were already hoping for and rarely get. Answering well makes their own return better, so the incentive pulls towards a considered answer rather than towards a completed form.

The base is around 660 responses across 17 rooms, February to August 2026. That is a sample, not a panel. It was not drawn to stand for any wider population and we do not claim that it does.

People select themselves twice: into the room, by choosing to attend, and then into answering. Both selections run in the same direction, towards people who came to get something out of the event, so anyone reading the results should expect the answers to overstate appetite relative to the full membership.

We do not weight the data. A count is a count of who wrote it in these rooms, not a share of a profession, and we do not project it onto one. Rooms differ in size, sector, seniority and country, so pooling across all 17 of them hides variation that matters. Where a finding comes from one room, we say which room and how many people were in it.

Free text is coded by hand, which is a judgement rather than a measurement, and the coding frame has developed as the rooms have accumulated. The completion figure describes an intake inside a registration flow, which is a friendlier context than an unsolicited survey, and it should not be read as a survey response rate.

What it is good for: finding what a fixed list of options misses; comparing stated demand against stated supply inside one defined group; and producing hypotheses precise enough that a properly designed study can go and test them. Overstating any of that would be worse than useless to us, because this method is only worth anything if the numbers hold up when someone checks.