Failure is rarely technical
Post-mortems on failed ERP rollouts seldom point at the software. Most point at process, expectations and data — and the warning signs are visible before anyone signs.
Here are the five most common, each with a question to ask beforehand.
1. Moving a broken process into the system unchanged
The most common and the most expensive.
A system freezes a process in place. If the current process is "sales messages the warehouse and the warehouse confirms verbally", putting it into an ERP just produces a more awkward messenger.
Ask yourself first: which three processes do we intend to change while we are doing this? If the answer is "none, keep it as it is", save the money.
2. Going live on uncleaned data
Duplicate SKUs, the same customer recorded three times, units written as "pcs" half the time, stock counts from three years ago.
In the spreadsheet era people worked around this from memory. A system will not. On day one, every workaround becomes an error message or a wrong number.
This task is routinely underestimated by three to five times. There is no shortcut, and it can only be done by people who understand the business — it cannot be outsourced.
Ask first: do our SKUs follow a unique rule? Is the customer master duplicated? When was the last physical count?
3. Only the owner wants it
The people deciding and the people typing are not the same group. The owner sees better reports; the operator experiences three fields becoming eight.
If nobody explains what those extra fields buy, operators will find a way around the system — and you are back at reason one.
Ask first: were the daily users involved in the selection? What do they hate most about the current process, and does the new system fix it?
4. Scope keeps growing mid-rollout
"While we are at it, let us add production scheduling."
Every added module grows testing, training and go-live risk — while the go-live date usually does not move. The result is every module half-finished.
A steadier approach: freeze the scope in writing and put everything else on a phase-two list. Phase two starts after phase one has run cleanly for three months.
5. Nobody defined success
After go-live, how will you know it worked? If the answer is "it feels smoother", there will never be a conclusion, and no basis for deciding whether to invest further.
Success should be measurable:
- Month-end close drops from three days to one
- Book stock value is within 2% of physical count
- Per-item margin is available in five minutes without manual calculation
- Channel stock no longer needs manual reconciliation
Write three sentences like these before signing. Without numbers, acceptance becomes an argument.
How we work
We ask all five before quoting, for two reasons.
First, some projects should not go ahead — adopting a system before the process is settled hurts both sides. Second, the answers make scope concrete, so the quote is not guesswork.
So if you ask us to scope something, expect the first conversation to produce questions rather than a price.