The seven questions that predict ERP implementation failure

I have been on both sides of an ERP selection — buying twice, and now building one. The pattern that stands out is how little the eventual outcome correlates with the evaluation criteria most organisations use.

Feature matrices are close to useless. Every serious ERP does general ledger, inventory and procurement. The differences that decide success are organisational, and they are visible before you sign anything if you ask the right questions.

1. Who owns the chart of accounts?

If nobody can name a person, the programme is already in trouble. The chart of accounts is the single most consequential design decision in an ERP implementation and the one most often delegated to whoever is available. It should be owned by someone senior enough to say no to a departmental request for another cost centre.

2. What does month-end look like today, in hours?

Not "how long does it take" — how many person-hours, across which teams. Programmes that cannot answer this cannot measure whether the ERP helped. It is also the single best proxy for how much undocumented process is about to be discovered.

3. Which system is authoritative when two disagree?

Every organisation has at least one pair of systems that report different numbers for the same thing, and a person who knows which to believe. If that knowledge lives only in someone's head, migration will surface it painfully. Write it down first.

4. How much of the current process exists because of a system limitation?

A large fraction of "how we do things" is scar tissue around a constraint in software being replaced. Teams that have not separated genuine business requirements from workarounds will faithfully reimplement the workarounds, at considerable cost.

5. Who is being freed up to work on this?

Not "who is involved" — who has been taken off their day job. An implementation staffed entirely by people doing it alongside their existing workload is a schedule that will slip, and everyone knows it except the steering committee.

6. What is the first thing that will go live?

The answer should be small, specific, and scheduled within about ninety days. If the first go-live is the whole system in twelve months, there is no feedback loop, no early evidence, and no way to correct course before the budget is committed.

7. What happens if we stop after phase one?

A programme where stopping early leaves you worse off than not starting is badly sequenced. Each phase should leave the organisation in a better state than before, so that a change in funding or priorities does not strand you mid-migration.

Why this works

None of these questions is about software. That is the point. In the implementations I have watched fail, the software was fine. What was missing was a named owner, a sequenced plan and a first increment small enough to actually finish.

Ask the seven questions before you shortlist. If the answers are uncomfortable, that is useful information, and it is far cheaper to learn now.