Counting association partner registrations without double paying for the same members
Association partner registrations should be counted against the overlap between the association's membership and your existing database. Match on email and employer domain before the campaign, then report registrations that were new to your file separately from those you already held and could have reached yourself.
The association partnership delivered 412 registrations. The commercial team is pleased, the association wants a bigger package next year, and nobody in the room can say how many of those 412 people were already sitting in your database with a valid email address and a history of attending.
Association partner registrations are the hardest partner number in the plan to read honestly, because the audience an association reaches is, almost by definition, the audience you have spent five editions collecting. The two lists overlap, and the overlap is where the money leaks.
Why the gross number overstates the deal
An association exists to gather the same professionals your show gathers. When one promotes the other, the promotion lands on people who are members because they work in the sector, which is the same reason they are in your file.
So a registration that arrives through the association's link tells you one of two things. Either the association reached somebody you could not reach, which is genuine reach you paid for. Or the association reached somebody you would have reached with next Tuesday's mailing, which is a registration you paid for twice: once in the partnership and once in the cost of running your own database.
Both look identical in a channel report. Separating them is the whole job, and it has to be done before the deal is signed rather than after the invoice.
How much overlap should you expect before you sign?
There is a back of the envelope calculation that takes ten minutes and beats any assumption.
Estimate the size of the addressable population for your sector. Say the industry your show serves has roughly 130,000 identifiable professionals in the markets you sell to. Your file holds 84,000 of them, which is 64.6 per cent coverage. The association has 6,000 members.
If membership were unrelated to being in your file, you would expect 6,000 multiplied by 0.646, which is 3,877 members already in your database, and about 2,123 new. That independence assumption is the neutral case, and it is usually pessimistic in the organiser's favour, because association members skew towards the engaged end of the sector and the engaged end is over-represented in your file already.
So the honest prior before any matching is that more than three fifths of a partner association's list is already yours. Any commercial conversation that assumes 6,000 new people is starting five thousand records away from reality.
Run the same arithmetic in reverse to sanity-check your own coverage claim. If your file holds 84,000 records against a population of 130,000 and a partner's matching exercise comes back showing only 20 per cent overlap, either your coverage figure is inflated by dead records or the association's membership sits in a segment you genuinely do not touch. Both are worth knowing and only one of them is good news.
Working the overlap on a 6,000 member list
Where the association can share records, the measurement is direct. Take their 6,000 and match against your 84,000 in two passes.
The first pass is normalised email: trim, lowercase, hash both sides. Suppose it returns 2,760 matches, or 46.0 per cent.
The second pass takes the 3,240 unmatched members and compares employer domain plus surname. Suppose 640 of those match a record you hold under a different address, usually a personal address on your side or a role address on theirs. Total overlap is 3,400, which is 56.7 per cent, leaving 2,600 members new to your file.
That 56.7 per cent sits below the 64.6 per cent independence baseline, which is mildly encouraging: this association reaches slightly outside your existing coverage. A partner whose measured overlap comes in at 80 per cent is selling you access to your own database.
Now price it. If the partnership costs 8,000 in cash or in reciprocal value, the cost per member reached is 1.33. The cost per new-to-file member reached is 8,000 divided by 2,600, which is 3.08. Neither number is a cost per registration yet, because reach is not registration, and the conversion from one to the other is what the campaign measures.
Suppose the campaign then produces 412 registrations carrying the association's code or tagged link. Of those, 176 are people who were not in your file before the campaign started, and 236 were. The cost per registration on the gross count is 19.42. On the new-to-file count it is 45.45. Both are true, and the second is the one that describes what the partnership added.
There is a third figure worth pulling out. If 34 of the 236 had already registered before the association's mailing went out, they were never available to be recruited, and the honest denominator drops to 378. That subtraction is small in this example and it is not always small, particularly when a partner mails late in the campaign.
Measuring overlap when you never see the list
The common case is that the association will not share member records, which is a reasonable position and often the correct one.
Two routes still work. The first is aggregate domain matching. Ask for the count of members by employer domain, with no names and no addresses. A count of 43 members at a domain is not personal data, and most associations will provide it because it costs them nothing. Match those domains against your own file and you can estimate overlap at the company level, which is a good proxy in a sector where employers are concentrated.
The second is a post-hoc split on the registrations themselves. Every registration carrying the association's code gets checked against a snapshot of your file taken before the campaign started, and only before, because a registration adds the person to your database and a check run afterwards will find everybody. Take the snapshot on the day the association mails and keep it. This is the single most common mistake in partner reporting, and it silently turns every new-to-file measurement into zero.
What you do with the people you already held is a separate question with its own taxonomy, and the version applied to exhibitor invitation programmes in counting exhibitor driven registrations transfers directly to association deals.
Who is allowed to send the mailing?
The mechanics of consent decide the shape of the deal, so settle them before the measurement design.
The normal arrangement is that the association mails its own members and you never touch the list. Guidance from the Information Commissioner's Office on electronic mail marketing, as published in 2026, is clear on why. The soft opt-in that lets you email your own customers about similar products does not extend to prospective customers or to contacts from bought-in lists. It is also explicit that you must comply with the rules if you send a marketing message or if you instigate somebody else to send it, so asking an association to promote your show does not move the responsibility entirely onto them.
The same guidance draws the line most organisers actually operate on: you can email a corporate body without consent, while sole traders and some partnerships are treated as individuals and need it. In an exhibition audience full of one-person consultancies and small dealerships, that distinction is not academic.
The practical consequence for measurement is that your tracking has to survive a send you do not control. Give the association a finished URL with the campaign parameters already in it and a code unique to them, exactly as you would for a trade title's placement, and ask them not to shorten the link.
Where this stops
The overlap number ages badly. A match run in January describes the file as it was in January, and by the time the show opens your database has grown, so re-running the same match will show more overlap and less partner value. Fix the snapshot date and cite it every time the number is used.
The second limit is that overlap in your file is not the same as reachability. A person can be in your database and unreachable: hard bounced, unsubscribed, or holding an address from an employer they left two years ago. Some of the 236 already-held registrations came from people you could not have mailed, which means the partnership recovered them, and only a check against your suppression and bounce state will show how many.
The third is that none of this measures whether the association's endorsement changed anybody's mind. Reach and overlap are countable. Influence is not, and a partnership with a respected association may be worth carrying at a poor cost per registration for reasons that belong in a different conversation, particularly when the exchange is reciprocal and has no cash in it at all, which is the subject of valuing a partnership where no money changes hands.
Context helps here. UFI's Global Exhibition Barometer, 36th edition, published in January 2026 from 378 companies across 57 countries and regions and produced in collaboration with 33 associations, reported 47 per cent of respondents seeing activity rise by more than 5 per cent in their domestic market in 2025. Associations are woven into how this industry measures itself, and a partnership discussion that treats them purely as a media buy misreads the relationship.
Start this week by taking one association partnership from the last edition and checking its registrations against a snapshot of your file dated before that campaign ran. If no such snapshot exists, that is the first thing to build, because every future partner measurement in your acquisition and attribution reporting depends on having one.
Questions people ask about association partner registrations
- How do you measure overlap with an association's member list?
- Match on normalised email first, then on employer domain plus surname for the residual. Where the association cannot share records, ask for aggregate counts of members by employer domain, which is not personal data, and match those against the domains in your own file. Both routes give a defensible overlap estimate before terms are agreed.
- Should association partner registrations count towards the marketing target?
- Count them, and publish the new to file share alongside the total. A partnership that produced 412 registrations of which 176 were people you did not hold has done something your own database could not. Reporting only the 412 credits the association with reaching people who were already on your mailing list.
- Can an association share its member list with an event organiser?
- Usually not, and the normal arrangement is that the association mails its own members on your behalf. UK guidance from the Information Commissioner's Office is explicit that the soft opt-in does not extend to contacts from bought-in lists, and that a sender is responsible when it instigates somebody else to send a marketing message.
Related reading
- Media partner promo codes that actually tell you which title delivered
- Valuing a media partnership when no money changes hands
- Measuring exhibitor driven registrations without taking credit away from marketing