An underused customer data platform is usually not a bad platform. It is a platform bought to solve a data problem when the real problem was a decision problem. Work out which one you have before you look at replacements, because changing vendors fixes the first and does nothing at all for the second.

Most underused platforms are the second kind, which is why the second platform tends to go the same way as the first.

This is worth saying plainly because the situation is common and people tend to carry it as a private embarrassment. A platform was bought with a business case, the business case has not landed, and somebody is going to ask about it. The useful move is a clear diagnosis, and the diagnosis is not complicated.

Tell the two apart

A platform problem looks like data that never arrived. Sources that were never connected, an identity match that does not work well enough to trust, events firing into a system nobody reconciled. The symptom is that the answers are wrong or missing.

A decision problem looks different. The data is there, the segments build, the dashboard loads, and nothing downstream changes because of any of it. Nobody was ever going to act on the output, because no one agreed in advance what they would do differently once they had it.

Why replacement feels right and often is not

Replacement is appealing because it is legible. There is a vendor to blame, a procurement process to run, and a visible decision to point at. It also resets the clock on the question everyone is uncomfortable with, which is what this was supposed to change.

The cost is rarely the licence. It is the integration work, the identity logic, the historical data that has to be migrated or abandoned, and the eighteen months during which nobody asks whether it is working because it is obviously still being implemented. If the original problem was that nobody acted on the output, all of that spending buys a better place to not act.

The option people skip

Pick one decision and make it work end to end. One audience, one action, one measurable outcome, on the platform already paid for. Not a strategy, one decision.

Two things come out of that. You find out quickly whether the platform can actually support a working decision, which is the question the replacement debate is really about. And if it can, you have a live example to argue from, which is worth more than another business case.

What to check before renewal

Whatever you decide, do the checking before the renewal date and not after it. The things worth knowing are whether the customer record can be exported in a usable form, whether the identity logic is documented anywhere outside the vendor, and how much notice the contract requires.

Those answers take an afternoon to gather while you still have a choice, and they are considerably harder to act on once the renewal has auto-extended. The removability test walks through the same questions in a scored form if it helps to have the list.

Where this comes from. These are the patterns behind underused platforms in Australian mid-market companies, written from delivery work and from conversations with people who inherited a decision somebody else made. Published by NTWRK, an Australian data and AI consultancy that has a commercial interest in the second branch of the diagram above. Worth reading with that in mind, and worth noting that the first branch genuinely ends at replacement. Related reading sits at how to choose a data consultancy.