A gearbox maker outside Pune: three plants, around 400 people, a little over ₹140 crore. The MD's daughter had taken over operations eighteen months earlier and had done the sensible thing. Five vendors, five demos, five proposals in a folder. When we met she had a problem I run into constantly. Every demo had gone well. Every vendor had said yes to everything. And she had no way to tell them apart.
The folder wasn't the problem. The scoring was. She'd been comparing feature lists, and a feature list is the one artefact every vendor optimises for. What actually separates these systems only shows up when you ask about the part that happens after the signature.
How should a manufacturer evaluate ERP systems?
The five weights, and why features aren't one of them
Capability in the abstract is the wrong axis. The most capable system on your shortlist is almost certainly the one you'll use least of, and you'll pay for the rest of it every year. What you're really buying is a fit, an owner, and a date. Weight accordingly.
| Criterion | Weight | What a good answer sounds like |
|---|---|---|
| Fit to how your plant runs | 30% | They describe your job work flow back to you before you finish explaining it |
| Who owns it after go-live | 25% | One named company, one named person, and a number you can call in month four |
| Weeks to first module in daily use | 20% | A date, with the module named, not a project plan with eleven phases |
| Separate bills you receive | 15% | They can list them without checking, including the ones that arrive later |
| Whether operators use it unprompted | 10% | They ask to talk to your supervisors before they quote |
Questions that separate one demo from another
Most demo questions get the same answer from everyone, which makes them useless for scoring. These don't.
- Show me a job card going out to a job worker and coming back short. Not a slide, the actual screen.
- Who does the implementation, and are they your employee or someone else's?
- Which module will be live first, and on what date?
- List every invoice we'll receive in the first three years, including the ones that only arrive if something changes.
- What happens in month four when something breaks and my supervisor is the one who notices?
- Which of your customers is closest to us in manufacturing mode, not in industry, and can I call them?
- What did your last rollout at our size get wrong?
- Show me multi-GSTIN despatch and an e-way bill, end to end.
- If we want a report you don't have, is that configuration or is that code, and who maintains it afterwards?
- What would make you tell us to buy something else?
Question six is the one people skip, and it's the highest-yield question on the list. Industry is a weak proxy. A pharma plant and a food plant share more machinery logic than two "engineering" companies where one makes batches and the other makes one machine over fourteen weeks.
Three answers that should end a call
The Pune evaluation ended two of its five conversations on these, and both vendors had scored well on features.
"We can build that." Asked about a gap, a vendor who reaches for custom code is telling you what your second year looks like. Custom code is billable to write and billable forever after, and it's the line item that turns a predictable subscription into an open-ended relationship. Configuration is a different answer and it sounds different.
"That depends on the partner." Fine as a fact, disqualifying as an answer to "who is accountable". If the product company and the implementing company are different, you have two relationships and one of them will point at the other in month four. We've written separately about whether you really need an ERP implementation partner.
Silence on question ten. A vendor who can't name a single situation where you should buy elsewhere either doesn't know their product's edges or isn't telling you. Both are expensive later.
"Four of them told me what the software could do. One of them told me what it couldn't. I bought from the one who told me what it couldn't, and I still think that was the whole decision."Gearbox manufacturer, Pune
Scoring it without pretending it's arithmetic
Score each vendor out of ten on the five criteria, multiply by the weight, and add it up. The number isn't the point. The point is that the number forces the committee to disagree out loud before the contract rather than after it, and in most evaluations the finance seat and the plant seat are scoring completely different things without having noticed.
Cut to three before you start scoring in detail. Five demos produce a folder; three produce a decision. If you want the field mapped before you pick the three, our ranked view of the best manufacturing ERP software in India covers nine systems and names the factory each one suits, including the ones that aren't us.
An evaluation that ends in a no still pays for itself, because the specification survives it. PPS International in Greater Noida had written its requirement down for a proposal it never signed, and went live elsewhere in eight weeks on the strength of it.
When you run this on Facto, question ten has a real answer: if you have several legal entities and a genuine statutory consolidation requirement, we're not your best buy and we'll say so on the call. If the shortlist is down to three and you want us in it, talk to our team and we'll take the questions above in order.
