Evidence

What membership organisations actually know about their own members

This page collects what the published research says about how well a chamber, an association or a professional body can substantiate what its members need. Every source on it is external, linked and dated. None of it comes from a company that sells software into this market, including us. Take any of it and use it - that is what it is here for.

Five claims follow. Each one states the claim as a sentence, sets out the evidence underneath it, and then names the decision it ought to change - because a number that changes no decision is trivia. After the five claims there is a separate section about our own data, kept apart from the published research and labelled as ours, with its limits written next to it rather than at the bottom.

01

You represent an industry on an educated guess, and the people funding you have started asking.

Only 11% of associations describe their own value proposition as ‘very compelling’. Half report no growth or a decline. It is difficult to hold both of those and still believe the sector knows precisely what its members want ASAE, 2025.

The pressure is not abstract. Dues have fallen from 95.7% of association revenue in 1953 to somewhere between 30 and 45% today ASAE, 2025. Everything filling that gap - sponsorship, training, accreditation, submissions, in some cases government-funded programmes - is sold on the strength of a population you claim to understand. The buyer of each one of those has learned to ask how you know.

11%

of associations call their own value proposition 'very compelling'

Half

report no growth or a decline in membership

95.7% to 30-45%

dues as a share of association revenue, 1953 against today

What this should change

Before the next board paper, write down the single sentence you would use to describe what your industry needs this year - then write down what it rests on. If the answer is a survey, a committee and a feeling, that is the finding.

02

The gap between what staff think and what members say has been measured, and it is large.

Across a study of 1,000 members and 200 association staff, staff rated one member benefit at 14% importance. Members rated the same benefit at 46%. That is a 32-point gap on a single item, between the people who run the organisation and the people it exists for Edge Research, 2025.

The uncomfortable part is not that the gap exists. It is that nobody inside the organisation felt it. A 32-point misread does not announce itself - it shows up as a programme that underperforms for reasons everyone agrees on afterwards.

32 points

between how association staff and association members rated the importance of the same member benefit, across 1,000 members and 200 staff

What this should change

Take your three most confident statements about what members value. Find the last piece of member-sourced evidence for each. If any of the three is carried by staff consensus alone, that is the one to test first.

03

Your member survey only hears from the members who were staying anyway.

The median response rate for a business-to-business survey is 8.18% Survicate, 2025. That is the middle of the distribution, not the bad end of it. Nine in ten of the people you asked did not answer, and the ones who did were disproportionately the engaged, the recently helped and the already committed.

This is not a failure of your survey design, and a better subject line will not fix it. Survey response has been falling across every mode for decades: telephone response fell from 36% in 1997 to about 6% by 2018 Pew Research Center, 2026. The instrument itself is degrading, and it degrades fastest at exactly the end of the membership you most need to hear from - the people who are drifting.

8.18%

median response rate for a business-to-business survey

36% to 6%

telephone survey response, 1997 against 2018

What this should change

Stop reading your response rate as a quality score and start reading it as a coverage problem. Ask who is missing from the answers rather than what the answers say, and put a number on how many members you have never heard from directly.

04

What you can evidence is what you can charge for.

When dues carried most of the organisation, understanding members was a matter of good service. Now that dues carry between 30 and 45% of it ASAE, 2025, understanding members is the product being sold. A sponsor is not buying a logo on a banner. They are buying access to a population you have described to them, and the description is the part they are actually paying for.

The market rate for that description, bought from a research firm, is instructive: business-to-business research projects typically run from US$18,000 to well over US$100,000 Adience, 2026. Organisations pay that because a credible account of a population is worth it. A membership body already sits on the raw material and mostly does not collect it.

Meanwhile the thing that keeps members - being understood - is the same thing that makes them stay. Median association renewal sits at 84%, and first-year renewal at 74 to 75% MGI, 2026. That first-year figure is the sound of people joining, not being reached, and leaving.

What this should change

Look at the last sponsorship or funding conversation that stalled. Ask whether it stalled on price or on proof. If it was proof, the thing to fix is not the deck.

05

The record you would own is publishable, and publishing it makes you the source in your sector.

63% of association leaders expect non-dues revenue to increase ASAE, 2025. Almost all of them are planning to get there through products. Very few are planning to get there by becoming the organisation everyone else has to cite.

A body that publishes a defensible annual account of what its industry needs stops being one voice among several and becomes the reference. That is a different kind of asset from a training course. It compounds, it cannot be undercut on price, and it makes every submission, every media call and every sponsor conversation start from a different place.

What this should change

Decide now who in your sector should be quoted when a journalist, a department or a rival body needs a number about your industry. If the honest answer is not you, nothing about that changes on its own.

Our own data - not published research

What we have found in our own rooms

Everything above is somebody else’s work. This section is ours, and it is a different kind of thing: smaller, newer, and gathered as a by-product of running introductions rather than as a designed study. We are setting it out here because it is the only part of this page nobody else can check, which means it needs its limits attached to it rather than mentioned at the end.

around 660 responses17 roomsFebruary to August 202660 to 70% of a room completes the intakeNot weighted

People ask for things nobody in the room can give them

The clearest thing we see is a two-sided one. When you ask a room both what people want help with and what they could help someone else with, the two lists do not match, and they fail to match in the same direction repeatedly.

A legal convention: what people wanted help with, against what anyone there could offer

One room. around 660 responses across 17 rooms in total, February to August 2026.

  • Wanted help with
  • Could offer help with
Legal practice
17wanted
7offered
10 more wanted it than could give it
Legal technology
11wanted
4offered
7 more wanted it than could give it
Data and technology
6wanted
0offered
Nobody in the room could help
The numbers
TopicWanted help withCould offer help withDifference
Legal practice177-10
Legal technology114-7
Data and technology60-6
Six people came looking for help with data and technology. Nobody in the room had it.Read as counts of people in one room, not as rates. A different room on a different day would produce different counts - the finding is the shape, not the numbers.

And the same question, on two topics where the imbalance runs the other way

  • Wanted help with
  • Could offer help with
Wellbeing
8wanted
18offered
10 more could give it than wanted it
Mentoring
17wanted
22offered
5 more could give it than wanted it
The numbers
TopicWanted help withCould offer help withDifference
Wellbeing818+10
Mentoring1722+5
Surplus is as useful a finding as shortage.Across four rooms the pattern held in both directions: growth and partnerships in deficit every time, personal effectiveness in surplus every time. That is a programming brief nobody had to commission.

The things people raise are mostly not on the form

Our intake offers about forty topics to choose from and a free-text box. Nine of the ten themes people wrote about in that box had no matching option among the forty. The form was not badly designed. It was designed by people who were not in the situation being described.

The five most common themes people wrote in that the form did not offer

Counts of people, around 660 responses across 17 rooms, February to August 2026.

Migrant and relocation experience
15
Burnout and recovery
13
Perfectionism
10
Imposter syndrome
9
Caring responsibilities
8
The numbers
ThemePeople
Migrant and relocation experience15
Burnout and recovery13
Perfectionism10
Imposter syndrome9
Caring responsibilities8
These are small counts and should be read as such.What matters is not that thirteen people mentioned burnout. It is that a well-intentioned list of forty options had no way for any of them to say so.

People answer, at a rate that surveys do not get

Between 60 and 70% of a room completes our intake. We are deliberately not calling that a response rate, because it is not the same instrument as a fielded survey - it is asked inside something the person had already decided to do, and answering it returns something they wanted. The comparison with 8.18% Survicate, 2025 is a difference of design, not of persuasion, and we would rather say so than let it be read as a boast.

Method

How the responses above were collected

Who was asked. People registering for an event, cohort or programme run by the organisation they belong to. The organisation had already decided who was eligible; we asked everyone who registered.

What was asked. A small number of short questions at the point of registration: what the person is working on, what they want help with, and what they could help someone else with. A list of about forty topics to select from, and a free-text box that is not optional in practice because most people use it.

What the respondent gets. A short list of named people in the same room worth finding, with a sentence on each saying why. That is the reason people answer, and we think naming it is more useful than implying the answers arrive out of goodwill.

What consent covers. People answer knowing the organisation running the event can see what they wrote, and knowing their answers are used to make introductions. Free text is read and grouped by hand.

What we do not do. We do not weight the responses. We do not reconcile a room against the membership it was drawn from. We do not treat counts from one room as rates for a sector, and we do not describe any of it as representative of anything, because people choose the room and then choose to answer - that is two layers of self-selection and no amount of framing removes them.

What would change our mind. The patterns above are consistent across the rooms we have run, but 17 rooms is 17 rooms. If a body ran this across its whole membership and the deficit pattern did not appear, we would want to know, and we would say so here.

Sources

Everything cited on this page, in the order it appears. Nothing here is published by a company selling software into this market.

  1. Marketing General Incorporated and GrowthZone, 2025 (via ASAE). The membership model is breaking down. Here is how associations can rebuild it. www.asaecenter.org/resources/articles/an_plus/2025/11-november/the-membership-model-is-breaking-down-heres-how-associations-can-rebuild-it
  2. Edge Research, 2025 (via Associations Now). Why associations need to invest in professional development. associationsnow.com/2025/10/why-associations-need-to-invest-in-professional-development/
  3. Marketing General Incorporated, 2026 (via Rallyboard). Retention rates: what associations can learn from subscription businesses. www.rallyboard.com/resources/blog/retention-rates-what-associations-can-learn-from-spotify-netflix-and-other-saas-companies
  4. Survicate, 2025. Survey response rate benchmarks. survicate.com/reports/survey-response-rate-benchmarks/
  5. Pew Research Center, 2026 (via SurveyMonkey). Survey response rate benchmarks. www.surveymonkey.com/learn/survey-best-practices/survey-response-rate-benchmarks/
  6. Adience, 2026. How much does business-to-business market research cost?. www.adience.com/blog/insights/b2b-market-research-price/
  7. Greenbook, no publication date shown. Online market research surveys: maximising your return. www.greenbook.org/marketing-research/online-market-research-surveys
  8. Veridata Insights, 2026. Incidence rate in market research: formulas and feasibility. veridatainsights.com/incidence-rate-market-research/
  9. Great Question, 2026. The complete guide to research incentives. greatquestion.co/blog/the-complete-guide-to-research-incentives

If you run a membership body and want a version of this about your own members rather than the sector in general, we run one room free to show you what it produces. Tell us about your organisation.