The default answer is buy. Standard ERPs are mature, someone else maintains them, and for a conventional manufacturing operation they are the cheaper and faster route. Custom ERP should have to earn its budget — but for a meaningful minority of Indian manufacturing SMEs, it clearly does. Here is how to tell which one you are.
The 80% test, applied to a plant
If a standard ERP fits 80% of your operation out of the box, buy it and adapt the remaining 20% of your process. If you are at 60% and pricing consultants to bend the product into shape, run the custom number honestly — customisation on a rigid product frequently costs more than a fitted build and leaves you owning nothing.
Test the 80% on the parts that are actually yours, not on the parts every business shares. Every ERP does purchase orders. Not every ERP does your job-work flow.
Where standard ERPs break in Indian manufacturing
These are the specific gaps we see repeatedly:
- Job-work and subcontracting. Material issued outside, partially returned, reconciled against a challan, with its own stock position and GST treatment. The most common breaking point.
- Multi-plant with shared inventory and inter-unit transfers priced differently from sales.
- Dealer and distributor hierarchies with regional pricing, schemes, and credit limits.
- Costing that reflects how you actually cost — by job, by batch, by machine hour, including rework and wastage.
- Industry-specific compliance where the format is dictated by a customer or an authority rather than by convention.
- Machine or shop-floor data capture from equipment that predates the internet.
If two or more of these are central to your operation, custom is a serious candidate. If none are, buy.
The comparison that matters
| Standard ERP | Custom ERP | |
|---|---|---|
| Upfront | Licence or subscription | ₹10L–₹35L |
| Recurring | Per user, forever | Hosting plus maintenance |
| Time to live | 2–5 months, mostly configuration | 4–9 months, phased |
| Fit | 60–90%, you adapt | Exact, if scoped properly |
| Ownership | None | Code, schema, and data |
| Upgrades | Vendor's roadmap | Yours, at your cost |
| Risk | Customisation debt, licence creep | Bad scoping, bad partner |
Compare three-year totals including planned headcount, not upfront numbers. And compare what you own at the end — a lower quote where the vendor hosts and holds the code is a subscription with a deposit.
The hybrid most plants should actually choose
Neither extreme is usually right. The arrangement that works most often:
Keep accounting where it is — Tally or your existing package, connected rather than replaced. Buy or build the standard modules — inventory, purchase, sales. Build custom only for the layer that is genuinely yours — job-work, costing, multi-plant transfers, or the dealer hierarchy. Integrate properly so nothing is re-typed anywhere.
This costs less than a monolithic custom build, survives better because commodity layers keep improving elsewhere, and it fails less often because each piece can go live independently.
Phase it, whichever you choose
- Inventory and purchase — the data foundation; nothing downstream is trustworthy without it
- Production and job cards — now costing becomes possible
- Costing and dispatch — where the commercial insight appears
- Integrations and reporting — last, on data you trust
Each phase must be live and genuinely used before the next begins. A phase that is "delivered but not used" is not delivered.
What decides success, and it is not the software
- A named internal owner with authority to decide. Without this, projects stall regardless of vendor quality — and this is on your side of the table.
- Clean opening data. Wrong opening stock on day one destroys trust permanently.
- Floor-level discovery. If nobody watched a shift, the system models an imaginary plant.
- A defined exception path. Supervisors bypass anything that blocks them mid-shift.
- Training that respects existing skill. Your supervisor knows the process better than the software does.
How to decide in two weeks
- Write the operation down, including subcontracting and the exceptions.
- Demo two standard ERPs against your own job-work and costing scenario, not their demo data.
- Note precisely where each breaks and what the workaround costs in hours.
- Get a written, itemised custom scope from a partner who scopes before quoting.
- Compare three-year totals, including licences, customisation, and headcount growth.
If it is close, buy. Custom should win clearly or not at all.
Get an honest read on your case
We build phased manufacturing systems — ERP development in Ujjain and ERP development in Indore, including the Pithampur belt. If a standard product fits you, we will name it and tell you to buy it.
For the smaller first steps, see software for manufacturers and business process automation.