Nuclear Institute

Where it started

The Nuclear Institute ran their membership on OMNET, the system that later evolved into Eudonet. When the time came to move, NI ran the decision internally and chose the new Eudonet platform on their own operational grounds. They didn't need us in the room, and we didn't have to weigh in on the decision. The website conversation came after, and that was our brief: make the web estate follow the decision, properly.

We mention the order because it matters, and because it's the right one. Your CRM is the operational heart of your association, and the choice of what runs it belongs to you, your team and your processes, not to a web agency's preferences. A website partner worth having doesn't need a vote. They need to be able to build well against whatever comes out of it.

NI told us the destination. We got to work.

The honest part

Then we met Eudonet's programming model, and we'll tell you plainly: we had never seen anything like it. Where most platforms present developers with entities by name, Eudonet speaks in tables, fields and catalogues identified by ID. It's the architecture that makes the platform so extendable, and it is genuinely disorienting the first time you meet it. It proved challenging at first, and anyone who tells you their first encounter with it wasn't is selling you something.

Two things got us through it fast, and both are worth naming because they're the real lesson of this project.

First, the Eudonet UK team were exceptionally supportive. They took the time to walk us through the programming model properly and the best approaches to take when integrating against it, rather than tossing us documentation and wishing us luck.

Second, Eudonet's full-featured Swagger endpoint meant we could prove every theory in the sandbox before committing a single line to real code. Every mapping, every workflow assumption, tested against a safe environment first. The learning curve was real, but it was navigated in a sandbox environment.

That's what de-risking an unfamiliar platform actually looks like. Not pretending the curve doesn't exist. Building the discipline that keeps mistakes in the sandbox instead of production.

The number we could have left out

We planned the integration at four months. It took nine.

Two things drove that, and neither was a mystery. The learning curve you've just read about was real, and we were climbing it for the first time. And Eudonet were building and customising NI's CRM while we were integrating against it, which meant parts of what we were connecting to were still taking shape as we connected to them.

We're publishing the number because it has become the most useful thing on this page. If you're planning a CRM change with a website integration attached, it carries two lessons.

First, sequencing matters. An integration against a CRM that's still being configured moves at the speed of the slower workstream. That's not a failure of either team, it's physics. Plan the overlap deliberately, or discover it expensively.

Second, the curve is paid once. Those nine months included the full tuition for Eudonet's programming model: the false starts, the sandbox experiments, the walkthroughs with the Eudonet UK team. That tuition was ours to pay, and it doesn't need paying twice.