Picture a distributor in Industrial Area. Eight months ago they bought a proper ERP to replace the spreadsheet-and-QuickBooks setup that was holding the business together. The licences are paid. The consultants came, configured, demoed, and left. And finance is still closing the month in Excel, because nobody quite trusts the numbers in the new system.

That story is common, and it almost never has anything to do with the software being bad. Microsoft Dynamics 365 Business Central, Odoo, Sage, SAP — the products work. ERP projects in Kenya fail for reasons that don’t show up in the proposal. After enough of them, you start to see the same handful of mistakes.

Nobody owned the data

This is the quiet killer. An ERP is only as good as what you put into it, and the data coming out of an old system is usually a mess — duplicate customers, stock items spelled three different ways, supplier balances nobody has reconciled since 2022.

Migration gets treated as an IT task to do at the end. It isn’t. It’s a business decision about what’s true. Who is the real list of active customers? What’s the actual opening stock? If a senior person in finance and operations doesn’t own those answers before go-live, the system launches full of garbage, people stop believing it, and they drift back to the spreadsheet they trust. Game over.

The organisation tried to keep every old habit

Every company has a way it has “always done things.” Some of it is genuine competitive advantage worth protecting. Most of it is just accumulated habit — workarounds for problems that no longer exist.

When a business insists the ERP must replicate all of it, two bad things happen. The implementation balloons with custom development that nobody really needs, and the standard, well-tested way the software wants to work gets buried under modifications that break on the next update. The skill in a good implementation is knowing which 10% of your process is special and worth building around — and having the discipline to adopt the standard for the other 90%.

”Go-live” was treated as the finish line

A surprising number of projects are scoped, priced, and staffed as if switching the system on is the end. It’s the middle. The first month after go-live is when people hit the edge cases, when a report doesn’t tie out, when someone in a branch does something the configuration didn’t anticipate.

If there’s no plan and no budget for that period — no one on call, no fast loop for fixes and small changes — confidence collapses exactly when it’s most fragile. We’ve inherited systems that were technically “live” and completely abandoned, because the support that should have carried users through week three was never part of the deal.

The team treated it as a software install, not a change

An ERP changes how people do their daily work. If the warehouse clerk, the accountant, and the branch manager weren’t involved early and trained properly, they will resist — not out of spite, but because the new way is slower for them until they learn it, and nobody helped them learn it. Training is not a one-day event at the end. It’s part of the design.

The boring things that actually work

None of the fixes are clever. That’s rather the point.

  • Put a real owner on the data, from the client side, with the authority to decide what’s true. Clean before you migrate, not after.
  • Adopt the standard wherever you can. Customise only where the process is genuinely yours. Every modification is something you’ll maintain forever.
  • Go live in stages where the business allows it — one module, one branch, one entity — instead of a single big-bang switchover that fails all at once.
  • Budget for the first 90 days after launch as seriously as the build. That’s when a project becomes either trusted or shelved.
  • Train the people who’ll use it, in their actual workflow, before and after go-live — not with a manual, with their own data.

This is most of what we mean when we talk about delivery certainty. It isn’t a secret methodology. It’s refusing to skip the unglamorous parts that decide whether a system goes live and stays live — which, with Business Central, is the whole job.

Thinking about an ERP, or trying to rescue one that stalled? Talk to us — we’ll give you an honest read on where it actually went wrong.