Do you need custom software, or will a spreadsheet do?
I build custom software for a living, so you would expect me to say you need it. Most of the time you do not. Here is how to tell the difference, including the point where a spreadsheet quietly starts costing you real money.
Spreadsheets are excellent, and that is the problem
A spreadsheet costs nothing, needs no training, changes in seconds and is understood by everyone. For a genuinely large share of business problems it is the correct answer, and anyone who tells you otherwise is selling something.
The difficulty is that spreadsheets do not fail loudly. They degrade. The file that was perfect at ten customers is merely awkward at eighty and quietly expensive at three hundred, and because each step was small, nobody ever decided to keep using it. It just kept being used.
When a spreadsheet is genuinely fine
Keep the spreadsheet if most of this is true.
- One person maintains it. No concurrent editing, no version confusion.
- The data is small. Hundreds of rows, not tens of thousands.
- The process changes often. If you are still working out how the thing should run, a spreadsheet's flexibility is an asset and software would lock in the wrong shape.
- Errors are cheap and visible. A wrong number gets noticed and corrected without consequence.
- Nobody outside needs access. No customers, no suppliers, no field staff.
If that is your situation, do not buy software. Spend the money on something that grows revenue.
The five signals that it has stopped being free
1. Two people cannot use it at once without a conversation
The moment there is a file named Stock_Final_v3_Rajesh_edited.xlsx, you are paying for the spreadsheet. Not in rupees, in reconciliation time and in decisions taken from the wrong copy. Cloud spreadsheets help and do not fully solve it, because two people editing the same logic still overwrite each other's intentions.
2. The same data is typed more than once
An order arrives, someone types it into the sales sheet, then into the stock sheet, then into the invoice, then into the accounts file. That is four opportunities for a typo and four places that can disagree. Duplicate entry is the single most reliable indicator that a process has outgrown its tools.
3. A monthly report takes half a day
If someone spends three hours a month assembling numbers by copying between sheets, that is roughly 36 hours a year of work that a tool would do in the time it takes to load a page. At any reasonable value for that person's time, the arithmetic stops being close.
4. Only one person understands it
This is the risk nobody prices. A spreadsheet with nested formulas built over four years by one person is a dependency on that person. When they leave, or are simply on holiday during a crisis, you discover how much of your business logic lived in their head.
5. You have started avoiding questions
The clearest signal of all. When "which customers ordered twice last quarter but nothing since" gets answered with "that would take a while", you have stopped being able to interrogate your own business. The information exists and is not reachable, so decisions get made on instinct instead.
What custom software actually costs
Being direct about numbers, because vagueness here helps nobody. For a small business in India, a focused custom web tool typically starts around ₹30,000 and runs to a couple of lakh depending on scope. An Android app starts higher. A business website is less.
Set that against what the spreadsheet costs. Take the hours lost per month to duplicate entry, reconciliation and report assembly. Multiply by the hourly cost of whoever does it. Add something for the errors, the delayed decisions and the key person risk, even if you cannot price them precisely.
For most businesses at the point where the five signals are showing, the answer comes out at a few months of payback. If it does not, that is a real answer too, and the honest conclusion is to keep the spreadsheet.
The middle option people skip
It is not spreadsheet or custom build. Between them sits off the shelf software, and it is frequently the right choice.
If your problem is genuinely standard, accounting, payroll, basic invoicing, an existing product will be cheaper and better tested than anything built for you. Do not commission a custom accounting system. That problem is solved.
Custom makes sense when your process is genuinely yours. A workflow with steps nobody else has, terminology specific to your trade, a way of pricing that no product anticipates. Then a template forces you to change how you work to suit software, which is the wrong way round.
The honest test: if you can describe what you need and someone can name three products that do it, buy one of those.
If you do build, build small
The commonest failure is not building the wrong thing. It is building too much at once. A project that tries to replace every spreadsheet in the business in one go takes months, costs more than expected, and arrives just as the requirements have moved.
Take the single most painful spreadsheet. Replace that one. Use it for a month. You will discover two things: some features you were certain about turn out not to matter, and something you never mentioned turns out to be essential. Better to discover that on a small build than a large one.
Then extend it. Software that grows one proven piece at a time ends up fitting, because each piece was shaped by actual use.
Three things to insist on
- You own the code. Not a licence to use it while you keep paying. If your business process lives in someone else's platform, you are renting your own operations.
- Your data is exportable in a plain format, on demand, without asking. If you cannot get your data out, you cannot leave, and you should always be able to leave.
- You see something working within two weeks. Not a design document. Something you can click. Any process that keeps you waiting months before you see the thing is hiding a problem.
The summary
Keep the spreadsheet if one person maintains it, the data is small, the process is still changing and nobody outside needs it. Buy off the shelf if your problem is standard. Build custom when the process is genuinely yours and you have counted the hours the current approach costs.
And if you are unsure, count the hours first. Not the ones you imagine. Ask whoever does the reconciliation how long it takes them. The number is usually larger than management thinks, and it makes the decision for you.
If you want a look at what a custom tool of this kind actually looks like, there are five working demos on the client work page that you can open and use.
Common questions
How do I know when a spreadsheet has become a problem?
Five signals. Two people cannot edit it at once, the same data gets typed more than once, a monthly report takes hours to assemble, only one person understands the formulas, and you have started avoiding questions because the answer would take too long. Two or more of those showing means it is costing you.
What does custom business software cost in India?
A focused custom web tool for a small business typically starts around ₹30,000 and runs to a couple of lakh depending on scope. Android apps start higher. Compare that against the hours currently lost to duplicate entry and reconciliation, priced at the cost of whoever does that work.
Should I buy off the shelf software instead?
Often yes. If your problem is standard, accounting, payroll, basic invoicing, an existing product will be cheaper and better tested. Custom is worth it when the process is genuinely specific to how you work, so a template would force you to change your operations to suit the software.
Why build in small pieces rather than all at once?
Because requirements you are certain about often turn out not to matter, and something you never mentioned turns out to be essential. Discovering that on a small build is cheap. Discovering it on a six month project is not. Replace the most painful spreadsheet first, use it for a month, then extend.
What should I insist on from a developer?
That you own the code outright rather than licensing it, that your data can be exported in a plain format whenever you ask, and that you see something working within about two weeks rather than a design document. A process that keeps you waiting months before you can click anything is hiding a problem.
Written by Khanjan Kavani, who builds software in Surat and writes these to answer the questions clients keep asking. Found a mistake? Tell me at hello@khanjankavani.com and I will correct the article itself.
General information, not professional advice. Rules in India change. Check anything important against the current position or a qualified professional. See the disclaimer.