MindT
Guide · Updated August 2026

Custom software vs Tally vs off-the-shelf ERP

We build custom software for a living, so you would expect this page to tell you to buy custom software. It doesn't. Most businesses that ask us for a custom ERP should not buy one, and we say so on the first call. Here is how to work out which side of that line you are on.

Start here: what is actually hurting?

Nobody wakes up wanting software. They wake up because a customer disputed a bill nobody can find, because stock on paper doesn't match stock in the godown, or because the owner cannot answer a simple question without three phone calls. Write down the specific thing that hurt you this week. That sentence decides everything below.

Side by side

What you needTallyOff-the-shelf ERPCustom
Accounting, GST returns, balance sheetPossible, but pointless — keep Tally
Standard purchase and sales billing
Your own process (lots, grades, pattis, commission)Partly, via workarounds
Staff using it on a phone in the field or yardSometimes, poorly
Works in Telugu / Hindi for non-accountant staffLimitedRare
WhatsApp bills and ledgers to customersAdd-onAdd-on
Changes when your business changesVendor's roadmap, not yours
Cost in year oneLowestLow to mediumHighest
Cost by year threeLowMedium — licences per user, per yearOne-time, then support only
Who owns the data and the codeYou own dataVendor owns the codeYou own both

Choose Tally alone if…

  • Your pain is accounting, GST and audit — not operations.
  • Your billing is standard: item, quantity, rate, tax.
  • One or two people do the data entry, in one place.
  • Your auditor is happy and your stock rarely surprises you.

This is most small businesses. Adding software here creates work rather than removing it.

Choose an off-the-shelf package if…

  • Your trade is common enough that packages exist specifically for it — pharmacy, garment retail, restaurants.
  • You can name three businesses in your exact line who use it and are still using it two years later.
  • You are willing to change your process to match the software, and your staff will accept that.
  • You have checked the total licence cost for five years, per user, not just the first invoice.

Choose custom only if…

  • Your process genuinely has no equivalent in a package — mandi lots and pattis, pond-wise aqua costing, milling yield, cold-storage chamber rent, dealer-specific pricing.
  • The daily workaround is costing real staff hours, and you can count them.
  • Your staff need it in Telugu or Hindi, on a phone, in a yard or a field.
  • You expect the process to keep changing and you want to control that, not file a request with a vendor.

MandiGrow exists because of exactly this. No package handled a commission agent's day — arrivals at 6:30 am, auction at 9, farmer payments after commission and advances, ledgers shared on WhatsApp by evening. So we built it, and it is now live in working mandis.

The arithmetic, done properly

Take the four questions below, write real numbers next to them, and the decision usually makes itself.

  1. 1How many staff-hours a day go into re-entering or reconciling the same information?
  2. 2What does one staff-hour cost you, fully loaded?
  3. 3What did disputes, wrong stock or missed dues cost you last year — an honest estimate, not a guess downward?
  4. 4How many years will you run this business the same way?

Hours × cost × 300 days, plus the losses, versus a one-time build plus support. If custom does not clearly win over three years, don't do it.

Red flags in any quote

  • A price given before anyone has watched how you work.
  • No fixed scope in writing — only an hourly rate.
  • No answer to 'who owns the source code?'
  • A demo full of dashboards but nothing that matches your actual bill format.
  • 'We'll do everything in three months' for a business with five departments.
  • Per-user licences that quietly grow as you hire.

Questions we get asked

Should I replace Tally?
Almost certainly not. Tally is very good at accounting, GST returns and your auditor's expectations. The right structure for most traders is Tally for accounts plus one system for operations, with data flowing between them so nobody types anything twice.
How do I know if custom is worth it?
Do the arithmetic honestly. Count the hours per day your staff spend re-entering, chasing or reconciling information, multiply by their cost, and multiply by 300 working days. If that number is a meaningful fraction of a one-time build, custom pays back inside two years. If it isn't, buy a package and stop.
What usually goes wrong with off-the-shelf ERP?
The implementation, not the software. You pay for customisation, staff find the screens confusing, they keep the real numbers in a notebook, and eighteen months later you are paying licence fees for something half-used. Ask any vendor how many of their customers in your exact trade are still live after two years.
Can we start small?
You should. Build or buy for the single most painful part of the day first — usually billing or stock — and only extend once that part is genuinely working. Anyone who insists on a full system in one go is protecting their invoice, not your business.
What if the developer disappears?
That is the real risk with custom, so make it a contract term: source code and database handed to you at delivery, documentation included, and no runtime licence key that can be switched off. We work that way as standard — you own everything.
Free, no obligation

Tell us what hurt this week

Twenty minutes on the phone. If Tally or a package is the right answer for you, we will say so — that call has ended that way plenty of times.