Every report told associations to do better. None measured what it costs to try.

I have spent twenty-five years building and running membership systems for Australian member organisations, and for most of those years I sold them the way the whole industry sold them. A core here, a website there, a module when you needed one, integrations to hold it together. What I have learned, from closer than almost anyone, is that the model that worked then does not work now.

Every few months another report lands on the desk of every association chief executive in the country. Engage more. Personalise everything. Prove value. Meet rising member expectations. Several arrived in the last year alone, from benchmark surveys to consultant blueprints, and together they read as a to-do list the length of your arm.

The advice is not wrong. That is worth saying plainly, because this is not an article about researchers missing the point. They have identified what members now expect with more evidence than the sector has ever had.

What nobody has done is measure the machine that has to do it.

There is no benchmark for how many systems hold part of one member's identity. None for how many staff hours disappear into reconciling those systems at renewal. None for what integration maintenance costs a mid-sized organisation, or how often a duplicate record turns into a wrong invoice, a wrong ballot paper or a wrong number in a board paper.

I read the current reports the sector is quoting this year. Between them they measure member sentiment, renewal rates, email performance, budget pressure and how confident staff are in their own data. Useful work, all of it. None of it puts a number on what the machinery costs to run.

We went looking for that number. It does not appear to exist. So we have started counting it ourselves, and if you run an association you can add yours in five minutes. the Association Systems Census

That absence is the subject of this article, because everything the reports are asking associations to do runs straight through it.

The leak everyone is politely ignoring

Start with the numbers the sector's own surveys report. Australian associations bring in new members at about 12.7 per cent a year and lose about 9 per cent out the back. That is growth, technically. It is also a bath with the plug half out.

Ask why members leave and the answer has hardened. Budget restraint was cited by 69.1 per cent of organisations in the 2025 Associations Forum survey, up nearly ten per cent on the year before, and it remained the leading reason in 2026. Members are weighing every dollar. The renewal decision is now a value test in a way it was not a decade ago.

Then the number almost nobody puts on a slide. First-year members renew at about 72 per cent, against a median of 82 per cent overall. That figure comes from an international benchmark rather than an Australian one, so treat it as a pattern rather than a local measurement, but the pattern is consistent everywhere it is measured. A quarter of the people you worked hardest to win leave before their second invoice.

Membership does not die in year ten. It dies in year one, in the gap between the promise made at joining and the experience delivered in the eleven months after. Not for everyone, obviously. But the first year is where the leak is widest, and it is the year most organisations watch least.

And many members do not leave suddenly. They leave slowly, and then officially. Long before the non-renewal arrives, the signals are already in the system. They stopped opening things. They stopped attending. They stopped logging in. They started renewing late instead of early. The non-renewal is not the decision. It is the paperwork for a decision made a year earlier.

Most organisations only measure the paperwork. Retention arrives as a single annual percentage, discussed at a board meeting after the members have already gone.

None of this is a motivation problem. Association teams work harder than they ever have. It is a machinery problem, and here is what that looks like from the inside.

Recruitment against attrition, Australian associations

Too many versions of the truth

I have seen this so many times over the years that it almost became normal, which is probably the problem.

An organisation thinks it has one membership database. It does not. There is the CRM. Finance has its own version. The website or portal has another. The email platform has another. And somewhere there is a spreadsheet somebody has been maintaining because they do not completely trust any of the others.

Then something important happens. A renewal, an election, a migration, an audit. Something where the answer to a very basic question suddenly matters: how many members do we actually have, and who are they?

That is when everyone discovers the numbers do not agree.

You export one list, export another, start matching email addresses and membership numbers, and there they are. Duplicates. Old records. People with two profiles. People marked financial in one system and unfinancial in another. People who changed employer or email address and, as far as the technology is concerned, became a different human being.

The interesting part was never the number. It was what happened in the room when it appeared, because at that moment it stopped being an IT problem. If you are running a ballot, who gets to vote? If it is renewals, who do you invoice? If you are reporting a membership figure to your board, how sure are you that it is right?

Then somebody asks which system is correct, and sometimes the honest answer is: we do not know yet. So staff who should be doing member work spend days comparing exports, ringing people, merging accounts and fixing statuses.

That is the part that stuck with me. Organisations can spend an enormous amount on software and still not have a dependable answer to the most basic question they have: who belongs to us?

There is a tell-tale artefact, and it turns up everywhere. Somebody builds a spreadsheet to bridge the gap between two systems, and quietly, that spreadsheet becomes the real system. The organisation is now running its most important asset on a file on somebody's desktop.

After seeing it enough times, I stopped thinking of duplicate records as a data-cleaning problem. It is an architecture problem. There are too many versions of the truth.

Worth doing right now, before you read on. Count the systems that hold a piece of one of your members: joining, renewals, payments, events, learning, email, the website, the spreadsheet nobody admits to. Write the number down. It is the subject of this article.

Renewal rate, first year against overall median

The reporting request that is never about reporting

Close cousin to the above, and one of the most common requests we hear: we need better reporting.

After twenty-five years I can say it is almost never true. When an organisation cannot answer a basic question about its own membership, the cause is rarely a missing report. It is that the underlying data cannot support the answer. Member categories that mean different things to different staff. Statuses applied inconsistently for years. Fields repurposed by someone who left in 2019. Three definitions of "active member" depending on who you ask.

A new reporting tool pointed at that data produces the same confusion, faster and in colour.

Definitions before dashboards. One agreed meaning for each status and category, one person who owns data quality, and the discipline to hold it. It is unglamorous and it is worth more than any reporting module ever built.

Renewal day, from the other side

Renewal day in most associations is nothing like the version the member experiences.

The member gets an email, clicks a button, pays. Behind that, half the organisation is holding the thing together.

Someone in membership is in early, checking the renewal population and making sure the right people carry the right fees. Finance has its own requirements, because eventually all of it has to reconcile with actual money. Communications needs a list, so there is an export. The portal sometimes needs another. Direct debit is its own process, failed payments are another list, and anyone who renews by invoice needs a human to make sure their status eventually catches up with the payment.

Then comes the part people outside these organisations never see: the reconciliation. Somebody with two spreadsheets open, checking who was invoiced against who paid, then checking the membership system to see whether the status changed, then working the exceptions. Why is this person still showing expired? Did we receive that payment? Which record is theirs? Why did this email bounce? Why has this organisation paid for nine people when there are ten on the account?

It goes for hours.

And the software has not technically failed. That is what makes it interesting. Every individual system is doing exactly what it was bought to do. The accounting system accounts. The email system emails. The CRM stores records. The gateway processes cards. The website serves pages.

The organisation is sitting between all of them, moving information from one place to another and reconciling the gaps. That is the hidden work. When something breaks at 7am on renewal morning, there is no layer that understands the whole process. A person becomes that layer. Somebody exports a file. Somebody fixes something. Somebody imports it somewhere else. Somebody checks the totals. Eventually somebody says: yep, they match.

After this many years, that always bothered me. We kept digitising individual tasks without ever really digitising the organisation.

Why the advice keeps failing on contact with reality

The most useful finding in this year's research is about members rather than machinery: those who feel their experience is personalised are markedly more likely to stay, and relevance now beats reach. Small, well-targeted sends outperform broadcasts by a wide margin, and the average association broadcast opens at about 33 per cent. The Associations Forum survey adds that many members already say they get too many emails.

Personalisation only works if one thing is already true. You have to know who the person is: their history, renewals, events, learning and conversations, held as one relationship rather than assembled on request from five places.

You cannot personalise from a stack that cannot agree who the member is. You cannot prove value per dollar when the evidence of value is scattered. You cannot fix first-year retention when no single view can see the whole first year.

There is a lazy version of this argument and I have no interest in making it. The lazy version says multiple systems are the problem and one system is the answer. That is not true and it is easy to disprove. Several specialist systems can hold one authoritative record perfectly well, when identity, ownership, synchronisation and failure handling have been designed rather than assumed. Some vendors are building exactly that, and they are right to.

The sector has actually settled the first question. Almost everyone now agrees there must be one authoritative record of each person. The argument has moved on to harder ground: where that record sits, who is accountable when two systems disagree, and who carries the cost of keeping it true.

That last one is the question nobody costs. In practice, in an organisation of six or ten people, the cost of keeping the truth true lands on staff. Not on a diagram, not on a vendor, not on an integration budget. On a person with two spreadsheets open on renewal morning.

Systems can be many. Accountability for the truth cannot. And somebody has to be paid to hold it.

Why this matters more now than it did ten years ago

The stack did not suddenly become bad. The job changed.

Four things converged. Members expect experiences that reflect their history with you, and the research ties that expectation to whether they stay. AI is only as useful as the identity and context it can safely reach, so contradictory records raise the cost and the risk of every AI initiative an association is about to be sold. Team sizes have not grown while expectations have; the sector's own 2026 survey says so, and points at smarter systems rather than more people. And participation now happens well outside financial membership, before it, alongside it and after it.

Put those together and the thing the system is organised around has to change. Membership systems were built with the membership as the organising object: a record, a status, a fee, and everything else hung off that. Associations are now asking them to understand a person and a continuing relationship, of which membership is one part. That is a different job. No amount of engaging harder fixes a machine that cannot say who the person is.

We know, because we built it

Bond Software has spent twenty years selling membership systems the way this market sold them. A core, then applications around it. Hosting. Integrations. Third-party components. Different suppliers looking after different layers. It was good software for what it was asked to do.

The model is a different question. It works right up until something has to cross a boundary. Then a diagram that looked perfectly sensible becomes a set of edges, and the organisation is the only thing holding them together. Nobody sold them that job. It arrived with the architecture.

The lesson I took was not build a better CRM. It was: stop making the customer carry the complexity of the architecture. If the organisation experiences something as one system, it should be designed as one system.

We did not read about the seams in a report. We built them, and then we spent years holding them shut.

There are really three ways to build this. The traditional membership system makes the membership the organising object and hangs everything else off it. The integrated stack keeps several specialist systems and asks them to coordinate identity between them, which works when it has been designed with real care and gets reconciled by hand when it has not. The third starts somewhere else: with the person. Membership is one thing that can be true of a person. So is being a learner, a sponsor, a supplier, an event attendee, a volunteer or a prospect, and several of those can be true at once.

That third one is the premise behind Nexy. We did not start with a membership record and build outward from it. We started with the person. A sponsor who later joins as a member does not become a second record; they keep the record they already had, and their history comes with them. Whoever someone is to you, and however many things they are at once, they are one person and the record knows it.

That sounds like a small distinction. In practice it is the whole thing.

Nexy is our attempt to build the platform we eventually came to believe the sector needed, rather than another version of the one we had spent twenty years building.

Which means I have a commercial interest in you agreeing with all of this. Discount it accordingly, then go and test it against your own organisation. That is the only check that counts.

What to demand from any platform vendor, including us

So what does this change when you next buy technology? Not the length of the requirements spreadsheet. Six commitments, in writing.

One authoritative record of each person. Not one system necessarily, but one place the truth lives, with named ownership of identity, status and the rules that govern them, and a documented answer to what happens when two systems disagree. If the demo shows three products "talking" and nobody can tell you which one wins an argument, you are buying the old problem in new packaging.

The whole commercial model before you commit. Licence, implementation, migration, every module you will actually use, integrations, and what happens at renewal. Complexity can legitimately affect price. Opacity should not. A vendor who will not put the shape of the commercial model in writing is telling you the price depends on you.

Implementation with its assumptions named. Not a date with no conditions, which nobody can honestly give. What is fixed, what varies, what it depends on from you, what the milestones are, and what happens if they slip.

Your data's exit route. How it comes out, in what format, at what cost, agreed before you sign. Honest vendors answer that in one paragraph.

Tell us what goes wrong with your product. Ask it straight. A vendor with a real answer has lived with their platform. A vendor with no answer is selling one.

A reason to stay put. Ask any vendor when you should not change platforms.

I have given that answer myself. There was a client where, looking at where they were, I said I do not think you should do this yet. Which is a strange conversation when you are the bloke meant to be selling them a system. But changing a core membership platform is disruptive: data, integrations, staff processes, finance, member communications, training. If an organisation is not ready to change how it operates, replacing the technology just moves the mess into a newer system. So the advice was: stay put, sort the data, get the organisation clear on what it actually wants, then come back to the technology.

The relationship did not disappear because we did not get the sale. In my experience the opposite happens when you are willing to say no, because they know you are not manufacturing a problem to suit what you sell. Sometimes the correct technology recommendation is: do not change the technology.

What comes after membership

Here is the part the sector's research circles without landing on. The financial member is becoming one point on a spectrum, not the whole audience.

Around every association already stands a larger crowd. The non-member who does your CPD. The follower of your advocacy. The event attendee who never joins. The lapsed member you could win back. The credential holder who wants the certification but not the badge.

What it looks like when it works is fairly unremarkable, which is the point. Someone follows your advocacy for a year without joining anything. They do a short course, and it lands on their record. They come to an event, and that is on the same record. When they finally join, nothing is re-keyed and nobody asks them for details they have already given three times. The organisation can see the whole path, and knows exactly what earned the membership.

Now ask what the installed machinery can do with that. Most systems can technically store a non-member. That is not the issue. The issue is what the system is organised around. Built to administer a membership and its transactions, the further a relationship sits from that transaction the thinner the record gets, and the more of it ends up in side lists and bolted-on tools. The fragmentation story again, this time with the organisation's growth attached.

The old organising question was: what membership does this record hold?

The new one is: who is this person, what is their relationship with us, what have they experienced, and what should happen next?

That change in what the system is organised around matters more than any feature anyone will show you in a demo.

The organisations that manage it will not be the biggest ones. They will be the ones whose systems can hold one honest record of every relationship, member or not, and do something with it.