"Should we buy SAP or build our own?" is the wrong first question. The right one is: how unusual is the way we actually operate? Companies that run standard processes should buy standard software. Companies whose competitive advantage lives in a non-standard workflow should not bend that workflow to fit a licence.
What each option is genuinely good at
Tally
Tally is excellent accounting software and a poor ERP, which is fine, because it was never meant to be one. It is unbeatable for statutory compliance, your CA already knows it, and it costs very little. It struggles once you need multi-location inventory, production planning, approval workflows or role-based access across departments.
Stay on Tally if your requirement is really books, GST and basic stock, and the pain you feel is Excel around the edges rather than Tally itself.
SAP Business One and similar mid-market suites
Genuinely capable, genuinely proven, and genuinely expensive once you count licences, the implementation partner, annual maintenance and the customisation you will inevitably need. You get decades of accumulated process wisdom — provided your processes look like the ones it encodes.
Buy it if you are scaling fast in a conventional manufacturing or distribution model, you need something auditable and internationally recognised, and you can staff the implementation properly.
A custom ERP
Built around how you already work, with only the modules you need, and no per-user licence that punishes you for growing. The trade-off is that you are commissioning software, so specification quality and vendor choice determine the outcome.
Build it if your workflow is your edge, if licence costs scale badly for your headcount, or if you have already spent two years and a lot of money customising a package into something it was not designed to be.
The cost comparison people get wrong
Licensed ERP is usually compared on year-one cost, where it often looks cheaper. The comparison that matters is five-year total cost, and it looks different:
| Cost line | Licensed suite | Custom build |
|---|---|---|
| Year-one outlay | Licence + implementation | Development |
| Per-user cost as you grow | Recurring, scales with headcount | None |
| Annual maintenance | Typically a % of licence | Support contract, negotiable |
| Changing a workflow | Partner change request | Backlog item |
| Ownership | You licence it | You own the source |
| Vendor lock-in risk | High | Moderate — depends on the contract |
Ask any custom vendor, in writing, who owns the source code and the database. If the answer is not "you, on final payment", walk away.
Phase it, or it will fail
The most common way a custom ERP project dies is a big-bang scope covering every department at once. It takes eighteen months, the requirements move underneath it, and nobody has seen a working screen. Sequence it instead:
- Phase 1 — Finance and Inventory. The transactional core everything else depends on. Go live here.
- Phase 2 — Sales and Procurement. Now that stock and ledgers are real, the order flow has something to write to.
- Phase 3 — HR, Payroll, Manufacturing. Layered on a system already carrying real data.
- Phase 4 — Analytics. Once there is a year of clean data worth analysing, add BI and dashboards.
Each phase should end with people actually using it. Parallel-running the old system for a month before cutover is not optional.
Requirements that are non-negotiable in India
- GST compliance — GSTR-1 and GSTR-3B exports, e-invoicing with IRN generation, e-way bills.
- TDS handling across vendor payments, with correct section mapping.
- Multi-state operations — separate GSTINs, inter-state stock transfers, place-of-supply rules.
- Tally or Excel migration — you have years of history and it has to come across cleanly.
- Mobile access for field staff, warehouse and approvals. Managers approve from phones, not desktops.
If any vendor treats these as change requests rather than baseline scope, adjust your budget expectations accordingly.
The shortcut: start from a product, customise from there
Custom-from-scratch is not the only alternative to a licence. For several common verticals a configurable product gets you eighty percent of the way on day one, and the remaining twenty percent is built around you. Our software products work this way — for example the School Management System for education groups and the Hospital Management System for healthcare.
Where no product fits, we build from scratch — see ERP development and custom CRM development, or look at what we have shipped.
A decision in one paragraph
If you are under about twenty people running standard trading processes, stay on Tally and fix the Excel around it. If you are scaling fast in a conventional model with budget for a proper implementation, evaluate a licensed suite seriously. If your operations are genuinely unusual, or you have already burned money customising a package into a shape it resists, build. And whichever you choose, phase the rollout.
Talk to us about your current setup — we will tell you honestly if you do not need a new ERP. Indicative pricing here.