When to replace a spreadsheet with a custom app

First: spreadsheets are underrated

A spreadsheet is fast to start, needs no permission from anyone, and every person in the business already knows how to use it. A great deal of real work runs on them for good reasons, and replacing one too early is as much a mistake as replacing it too late — you trade something everyone can change for something only a developer can.

So the question is not whether a spreadsheet is "proper software." It is whether this particular spreadsheet has started doing a job it was never shaped for.

The signs it has outgrown the job

• Several people edit it, and they collide. Files named final_v3_REAL, rows that change back overnight, a rule that only one person may touch column F.
• It is the only record of something the business depends on. The customer list, the job schedule, the inventory. If the file were deleted or corrupted tomorrow, what would you lose?
• Someone re-types data into it from somewhere else. Copying from emails, a form, or another system by hand is where most spreadsheet errors are born.
• The logic lives in one person's head. Formulas nobody else can read, and a quiet dread about that person's holiday.
• It gets emailed around. Every attachment is a fork, and nobody is sure which fork is true.
• Errors are found downstream. A customer notices the wrong price before anyone in the business does.
• You need people to see different things. Staff should see their jobs but not the margins. Spreadsheets are bad at "some rows, for some people."

What to try before custom software

Be honest about the cheaper fixes first. Several of the problems above have them:

• Move it to a shared, cloud spreadsheet with protected ranges. That alone ends the emailed-attachment problem.
• Put a form in front of it, so data arrives in a consistent shape instead of being typed straight into cells.
• Buy the off-the-shelf tool for the job if the job is common. A standard CRM, scheduling or inventory product is often exactly right, and buying is the right default.
• Automate the re-typing if that is the only real problem. Workflow automation can move data between systems without replacing any of them.

Custom software earns its place when none of those fit — usually because the process is specific to how your business works, and every product you have tried makes you bend the business to suit it.

What a replacement actually involves

Less than people fear, usually. The spreadsheet itself is the best requirements document you will ever have: its columns are the fields, its tabs are the screens, its formulas are the rules, and the workarounds people have built around it are the features it is missing.

A good replacement keeps what works about the spreadsheet — an export anyone can open, a layout people recognise — and removes what does not: collisions, lost history, re-typing, and the single person who understands it. The measure of success is not new features. It is that nobody has to be careful anymore.

Common questions

Is a no-code tool like Airtable a custom app?

It sits in between. A no-code database fixes several spreadsheet problems — structured records, permissions, forms — without writing software, and for many teams it is the right next step. It stops being enough when the logic outgrows what its rules can express, or when it has to talk to systems it has no connector for.

Can we keep using Excel after replacing the spreadsheet?

Usually, yes, and you should be able to. A replacement that cannot export to a spreadsheet takes away the one thing everyone already knows how to use. The difference is that the export becomes a report you read, not the record you edit.

Who should own the replacement?

You. The data, the code and the right to take both elsewhere should be yours in writing, whoever builds it. A replacement that locks the record inside a vendor has swapped one single point of failure for another.

Related