Why generic ERP fits Coimbatore units badly
Coimbatore runs on pumps and motors, foundries, engineering components, wet grinders, auto parts and textiles, with Tiruppur knitwear and Erode processing along the same corridor. Almost all of it involves material moving between your floor and somebody else machining shop, and almost no off-the-shelf ERP models that properly.
Generic ERP assumes you buy material, make a thing and ship it. A Coimbatore unit buys castings, sends them out for machining, gets them back short by two, sends a batch for plating, receives a rejection, and has to know at any moment how much of its own stock is sitting in somebody else premises. Packages either ignore that or bolt it on as an afterthought, which is why so many implementations end with a parallel spreadsheet that becomes the real system.
Job work and subcontracting, modelled properly
Outward job work needs a challan that names the material, the quantity, the process and the expected return date, and it needs to stay open against your stock ledger until the material comes back. Short returns, rejections and scrap have to reduce that balance explicitly rather than quietly. Inward job work - where you are the machining shop - needs the mirror of the same thing so the customer material never mixes into your own stock valuation.
Get that right and a question that currently takes a phone call and a ledger book becomes one screen: what is out, with whom, since when, and what is overdue. In practice this single module is the reason most Coimbatore units commission custom ERP at all.
The modules that actually earn their place
A staged build beats a big-bang rollout every time, because each stage goes live, gets used and pays for itself before the next begins. This is roughly the order that works.
| Module | What it replaces | Typical build |
|---|---|---|
| Item master & BOM | Part-number spreadsheets, drawings in a folder | Foundation - always first |
| Purchase & GRN | PO books, WhatsApp approvals, manual goods receipt | 2-3 weeks |
| Stores & stock ledger | Bin cards and a stock register | 2-3 weeks |
| Job work in / out | Challan books and a phone call to the vendor | 3-4 weeks |
| Production & stage tracking | Job cards on paper, supervisor memory | 3-4 weeks |
| Dispatch, invoice & e-way bill | Separate portal, re-typed invoice details | 3-4 weeks |
| Quality & rejection tracking | A register nobody reports from | 2 weeks |
| Reports & Tally export | Month-end spreadsheet assembly | 1-2 weeks |
Compliance built in, not bolted on
E-way bills are a daily operational task for a unit dispatching constantly, not an occasional chore. Raising them from the same record as the invoice - so the transporter, vehicle number, distance and HSN are already filled - removes the single most re-typed thing in the building. E-invoice IRN generation above the turnover threshold works the same way.
The accounts team works in Tally regardless of what the floor uses, so clean Tally-compatible export is a requirement rather than a feature. The test of a good ERP here is that nobody is typing the same number into two systems, and that is the specific thing to check in any demo, from me or anyone else.
Rollout without stopping production
ERP rollouts fail in the transition, not in the build. The approach that works is one module live at a time, with the old method running in parallel for a fortnight, and opening stock loaded as a deliberate cut-over with somebody accountable for the count.
Floor screens are built mobile-first and tested on the cheapest phone in the unit, because a supervisor logging a job card is standing next to a machine, not sitting at a desk. Where the team works in Tamil, the screens they use can be in Tamil - that alone decides adoption more often than any feature.
Buy a package or build one
Buy if your process is genuinely standard and you can adapt to the package. Odoo, TallyPrime with add-ons and the Indian manufacturing ERP vendors all work well for units whose flow is conventional, and they are cheaper and faster to start than anything custom.
Build when the job-work flow is the business, when the package forces a workaround that becomes the real process, or when per-seat licensing over a five-year horizon exceeds a one-off build. If a package fits, the scope document will say so - it is a worse outcome for me and a better one for you, and pretending otherwise is how consultants get a reputation.
Frequently asked questions
Can the ERP handle outward and inward job work?
Yes, and it is usually the reason the project exists. Outward challans stay open against your stock ledger until material returns, with short returns, rejections and scrap reducing the balance explicitly. Inward job work keeps customer material out of your own stock valuation.
Does it generate e-way bills and e-invoice IRN?
Yes, from the same record as the invoice, so transporter, vehicle, distance and HSN are already filled rather than re-typed into a separate portal. E-invoice IRN generation applies above the turnover threshold.
Will it export to Tally?
Yes. Tally-compatible export and GSTR-ready reports are standard, because the accounts team works in Tally whatever the floor uses. Removing double entry is the point of the integration.
How long does a manufacturing ERP take to build?
Staged, not in one go. Item master and BOM first, then purchase and stores, then job work, then production and dispatch. Each stage is two to four weeks and goes live before the next starts, so a useful subset is running within about two months.
Should we just buy Odoo or a packaged ERP instead?
If your flow is conventional, yes - a package is cheaper and faster, and the scope document will tell you so. Build when job work is the core of the business, when the package forces workarounds that become the real process, or when licence cost over five years exceeds a one-off build.
Can shop floor screens be in Tamil?
Yes. Floor-facing screens can run in Tamil while reports and admin stay in English, which is the usual arrangement. It affects adoption more than most features do.
What if our internet drops on the shop floor?
Entry screens can be built to tolerate it - queue the entry locally and sync when the connection returns - so a supervisor is never blocked. It is worth specifying up front, because retrofitting offline behaviour is considerably more work than designing for it.
Map your ERP in one call
Tell me what you need and you get a written scope, a fixed quote and a delivery timeline before any work starts — no obligation.
Ananth N · Madurai, Tamil Nadu · serving Madurai, Coimbatore, Chennai and clients across India · remote-first.
Request an ERP scope WhatsApp Email