When off‑the‑shelf integrations aren’t enough

If you’ve ever tried to connect your website, CMS, payments and CRM using an off-the-shelf integration tool — whether that’s Zapier, Make, Microsoft Power Automate, or the like — you’ll know the initial appeal.

It promises speed. It promises simplicity. It promises that your team can “just automate it”, no developers required! And for straightforward use cases like a single form, a single destination, a single happy path, that promise can hold up. But the organisations we work with rarely have straightforward use cases.

Large UK charities, cultural organisations, and leading visitor attractions don’t run on one form and one database table. They run on interconnected workflows: memberships, donations, events and ticketing, finance, marketing consent, Gift Aid, all across multiple teams, systems, and service providers.

In that world, “the integration” isn’t just a connection between systems, it's operational infrastructure.

And when it’s not done right, the failure is reflected in supporter dissatisfaction, in staff frustration, and often in the bottom line.

So how do you know if your integration setup is hurting you more than it's helping?

We believe the first warning sign will be background noise — the constant low‑level frustration in your team that you can ignore at first, but builds. Then before you know it, the CRM gets blamed for everything, because nobody can quite see where the problem is.

You might recognise it as:

• “The CRM is full of duplicates.” Not just occasionally, constantly.
• “We can't trust it.” Integrations randomly fail, partly run, or quietly drop data without telling anyone.
• “Can we update the record manually?” A quick fix that becomes the normal process.
• “Reporting is a nightmare.” The numbers only make sense after someone exports, filters, compares, and patches the gaps.
• “We don’t know what happened to your membership renewal.” Your team spends more time investigating technical issues than giving real value to members.
• “We’ll fix it later.” A growing backlog of workarounds and exceptions never quite get resolved, and slowly choke efficiency.

These issues all point to a setup that doesn’t hold together operationally.

When the CRM starts to feel like a place to tidy up rather than a system of record, everything slows down. Staff spend time compensating for gaps and exceptions instead of improving supporter journeys.

At that point, it might be time for a rethink.

In a complex organisation, the workflow rarely lives in one place. It spans your website and CMS, your CRM, your payments platform, email marketing, ticketing or membership systems, plus internal processes that have evolved over years.

Off the shelf tools may be able to handle the separate elements of this system perfectly well. But that's the core issue: by splitting everything into separate automations you end up with a patchwork system:

• multiple streams owned by different departments
• slightly different field mappings in different places
• gaps between systems that staff quietly fill in
• processes that only work with a workaround

That’s why we see these setups fail. Not because the tools are “wrong”, but because the organisation needs integration to behave like infrastructure: one joined-up system that can keep working through the complexity, through the edge cases, through the messy reality.

If your system isn't joined up, data loses meaning.

To spell it out clearly: if membership status, consent preferences, Gift Aid declarations, event bookings and attendance history, payment schedules and all the rest of it (take a breath — this is a long sentence) ... if all these data don’t arrive together, consistently, in the right structure, with the right rules, your CRM starts to fill with information you can’t confidently use.

When you’re dealing with operational complexity, you need more than a connector. You need a single integration capability that sits between systems, reduces fragmentation, and gives you one place to manage rules, reliability, and visibility.

That’s where bespoke middleware steps in.

It starts with a simple but fundamental goal: don’t let data disappear.

Join the queue (and get your data in line)

In our work with WWT, we built middleware to stop website submissions firing off into the void. Now they join a queue, get processed reliably in the background, and leave behind a clear record of what happened.

That means if something fails downstream, the team no longer rely on guesswork — they can see what was submitted, when it was submitted, and whether it successfully made it into the CRM.

We’ve taken the same approach with organisations like DACS, Chester Zoo, and Churches Conservation Trust where the middleware runs a stream of discrete jobs with clear statuses. If something fails, it’s visible, and crucially, recoverable. A staff-facing control panel can surface what failed, why it failed, and allow a safe retry at the right point rather than starting again from scratch.

APIs can be flaky, and that’s normal. Services occasionally slow down or fail, requests time out, tokens expire, and data arrive incomplete. What matters is creating a system that can recover safely without losing meaning or creating a backlog of manual fixes.

For WWT, these efforts turned a system where transactions failed nearly 30% of the time into one where errors dropped to less than 1%.

And then there’s the work most people underestimate: the small-but-critical data hygiene rules that stop everything slowly degrading over time. De-duping, name formatting, email validation, and correctly handling phone numbers are the kind of unglamorous details that emphatically stop the rot, and return the CRM from a degrading asset back into a system you can truly trust.

Every established organisation has quirks. Sometimes they’re legacy. Sometimes they’re inconvenient. Sometimes they’re the result of years of teams adapting to real-world constraints.

But quirks aren’t inherently a problem, they only become so if they’re dealt with using generic remedies or manual workarounds.

A well-designed middleware layer gives you a better option: it supports your quirks — sorry — your real-world processes end-to-end, with structure and reliability, without compromising data hygiene.

It also means you can build custom reporting that reflects how your organisation actually works, so fundraising, membership and comms decisions can be made confidently.

There’s no perfect test, but we see the same complaints time and again.

If you’re struggling to reconcile website activity with CRM records, if staff are spending time fixing data, or if your reporting is only accurate after someone’s manually cleaned it up, then plug-and-play automation probably isn’t enough.

At first glance middleware might seem like over-engineering. But because off-the-shelf tools are only great when your problem is simple, a custom integration layer may well be the simplest way to bake lasting value into your digital operations.

When your organisation is built on operational complexity, you need integration infrastructure that is resilient to API failure, designed for real workflows, supportive of long-standing business quirks, and built with reporting and governance in mind.

Because the goal isn’t just that data arrives.

It’s that the organisation can trust what happens next.