Stop rewriting your ERP. Integrate around it.
Every CFO has been pitched a rewrite. Almost none of them needed one. The real win is in the eight systems your ERP is supposed to talk to but doesn’t — yet.
Enterprise ERPs are unloved for a reason. The reports are dated. The UX is slow. The mobile experience is a desktop on a smaller screen. So the recurring fantasy — endlessly seeded by integrators selling implementations — is to rewrite. Rip out SAP. Rip out Sage. Rip out the home-grown thing the founders built in 2007.
In our experience, that rewrite almost never makes economic sense. And we have run the numbers on enough of them to say so without flinching.
A 2-year rewrite of a working enterprise ERP at a mid-market corporate in Zimbabwe lands somewhere between US$ 850k and US$ 2.4m, depending on customisation depth, integrations, training, and the inevitable parallel run. It also lands with a 60–70% chance of slipping by 6+ months — the industry track record is stubbornly bad.
For comparison: the integration-and-BI-on-top approach typically lands at US$ 80k–US$ 220k for the same scope. Two-month delivery, not two-year. And it leaves the ERP vendor’s support relationship intact — meaning when something breaks at 2am, you have a phone number that isn’t ours.
The reason the integration approach wins is geometric, not technical. Enterprises do not have an ERP problem. They have an integration problem. The ERP is doing what ERPs do — booking the journals, holding the master data, running the period close. The actual pain is that the ERP doesn’t talk cleanly to the CRM, the warehouse system, the billing platform, the field-service app, the finance dashboards, the HR system, or the BI tool. That is eight integration points. Fix six and the ERP suddenly looks fine.
The trap is that the ERP team and the BI team and the integrations team report to three different leaders, and the rewrite pitch unifies them all under a single vendor invoice. It is procurementally tidy. It is operationally a disaster.
“Most enterprise pain isn’t the ERP. It’s the eight systems the ERP is supposed to talk to but doesn’t.”
When we walk in the door, the first thing we do is map every system that touches the ERP and rate the integration maturity. Score 0 = manual export to Excel; 5 = real-time event-driven. We have never seen an enterprise where the average score across 8 systems is above 2.5. That tells you where the leverage lives.
There is a category of cases — typically 1 in 10 — where the rewrite is the right call. The ERP is unsupported. The vendor has gone bust. The customisations have produced a Frankenstein that no integrator wants to touch. Fine — rewrite. But it should be the last move, not the first.
The pitch we give CFOs: give us 4 weeks to map your integration landscape. We will come back with a written audit, a target architecture, and a phased plan that costs less than your last quarterly licence renewal. If at the end you still want to rewrite, you’ll be doing it from a position of clarity and we’ll walk away.
Most of the time, they don’t want to rewrite anymore.
More from the team
Why AI tutors will land harder in African classrooms than anywhere else
A 60-pupil classroom is not a bug to fix — it’s the constraint that makes a well-built AI tutor leverage instead of replace teachers. Here is how we are building one.
Read MarketingWhatsApp is the new front door — design for it like you mean it.
Half your customers are not coming through your website. They are coming through WhatsApp, and they expect the same product polish.
Read