Excel to web app: when to switch, what it costs and what breaks
Seven signs you've outgrown the spreadsheet, what research says about spreadsheet errors, the options in between, and what moving to a web app really costs.
Most businesses we meet in Indore run on at least one spreadsheet that has quietly become "the system": the stock sheet, the party ledger, the order tracker. That's not a mistake. Excel is the fastest way to start. The trouble starts when the business grows and the sheet doesn't.
This guide covers the signs that you've outgrown a spreadsheet, the realistic options in between (including when you shouldn't build anything), what migration really involves, and what it costs.
The short answer
- Stay on a spreadsheet if one or two people use it, the data fits on a screen or two, and a wrong number won't cost you money.
- Move to Google Sheets or Excel online if the problem is mostly "five versions on WhatsApp". Shared files with version history fix that for free or almost free.
- Try a no-code tool (AppSheet, Power Apps) if you need simple forms on phones and your data is small.
- Build a web app when several people enter data every day, mistakes cost real money, you need to know who changed what, or the sheet has become the bottleneck in how you take orders, bill or manage stock.
Spreadsheets go wrong more often than people think
This isn't about careless staff. Researchers have audited real business spreadsheets for decades, and the results are consistent:
- Professor Ray Panko's summary of field audits found errors in 88% of the 113 spreadsheets examined across seven studies (Panko, 2008). It's older research, but later audits haven't contradicted it.
- In commercial audits, roughly 1–2.5% of formula cells contained errors (Panko, 2015). On a sheet with 2,000 formulas, that's 20 to 50 wrong cells, and nobody knows which ones.
- When researchers studied the Enron email archive, 24% of spreadsheets with formulas contained an Excel error value like
#REF!or#DIV/0!, and about 1 in 10 emails sent or discussed a spreadsheet (Hermans & Murphy-Hill, 2015). Version chaos isn't new.
Excel can also change your data without telling you. A study of published scientific papers found that 30.9% of those sharing gene lists in Excel had names silently converted into dates (Abeysooriya et al., 2021, PLOS Computational Biology). Indian businesses see the same thing in different forms: PIN codes losing their leading zero, long account or GSTIN-like codes turning into 2.3E+14, and 04/05 being read as 4 May on one computer and 5 April on another.
Seven signs you've outgrown the sheet
- Several people edit it every day, and someone keeps a "master" copy that is never quite up to date.
- File names like
orders_FINAL_v3_REAL.xlsx. If people send the sheet on WhatsApp or email, you have several truths. - Nobody can tell who changed a number, or when. If you're a company keeping accounts in software, the law already expects an edit log: since 1 April 2023 the Companies (Accounts) Rules require accounting software that records an audit trail which can't be switched off (summary by Tally Solutions). A shared Excel ledger can't give you that. (The rule applies to companies, not proprietorships, but the risk is the same.)
- GST details live in formulas. Rates, HSN codes and invoice numbers typed or calculated by hand. If your turnover is ₹10 crore or more, e-invoices must now be reported within 30 days of the invoice date (GSTN advisory, 5 Nov 2024), so invoices batched up in a sheet now carry a deadline.
- WhatsApp is your backup. If the laptop dies, the business stops.
- One person understands the formulas. When they're on leave, nothing gets done properly.
- Reports take hours. Month-end means copying, pasting and checking totals by hand.
Three or more of these, and the sheet is costing you more than it saves.
You're not behind: where Indian businesses actually are
Most small businesses in India are still at this stage. A 2025 survey of 2,882 MSMEs across 14 states (7% of them in Madhya Pradesh) found that while 87% take digital payments, only 12% use ERP software and 14% use cloud tools (RIS / Mastercard, MSME Digitalisation in India). A separate vendor survey of 5,149 MSMEs by Zoho reported that 71% still use spreadsheets to manage customer information (Business Today, 2024).
In other words, the money moved online years ago. The operations mostly haven't.
Your options, honestly compared
| Google Sheets / Excel online | AppSheet (Google) | Power Apps (Microsoft) | Custom web app | |
|---|---|---|---|---|
| Best for | Shared lists, small teams | Simple phone forms on Sheets data | Teams already on Microsoft 365 | Daily operations: orders, stock, billing |
| Size limits | Sheets: 20 million cells per file. Excel: 1,048,576 rows | AppSheet database: 2,500 rows per table on Starter/Core | 250 MB database included; more at extra cost | Whatever your server holds |
| Cost (checked Oct 2026) | Free or included in Workspace / Microsoft 365 | $5–$20 per user per month | About $20 per user per month (Premium) | One-time build + hosting (see below) |
| Who changed what | Version history only | Basic | Yes, with setup | Full audit log, designed in |
| Lock-in | Low | High | High | Low if you own the code |
Sources: Excel limits, Google Sheets limits, AppSheet pricing, Power Apps pricing.
Two things the brochures don't stress:
- Excel co-authoring only works properly when the file sits in OneDrive or SharePoint with a Microsoft 365 subscription. A file on someone's desktop shared over WhatsApp is not co-authoring.
- Power Apps has a quiet trap called delegation. When a query can't be handed to the data source, Power Apps reads only the first 500 rows (10,000 at most) and works on those, so a search or total can be wrong without any error message. Excel files are one of the sources where this applies (Microsoft: delegation overview).
No-code tools are genuinely useful. A 2023 study in IEEE Access that surveyed experts on low-code found 12 drivers and 19 inhibitors, with lock-in, scaling and integration among the recurring concerns (Käss, Strahringer & Westner, IEEE Access, vol. 11, 2023, open access). Our rule of thumb: if the app is small, internal and unlikely to grow, no-code is fine. If it's the core of how you take orders or bill customers, own the software.
What actually breaks when you migrate
The building is rarely the hard part. These are:
- Dirty data. The same party typed three ways, merged cells, totals typed over formulas, dates in two formats. Expect to clean before you import.
- Rules hidden in formulas. A VLOOKUP that applies a special rate to one dealer, a column someone colour-codes by hand. Each one has to be found and written down, or the new system will "lose" it.
- Steps nobody wrote down. "Ramesh checks this every Monday and messages the godown." Software only does what it's told.
- The switch-over. Run the old sheet and the new app side by side for two to four weeks and compare the numbers before you stop using the sheet.
What it costs and how long it takes
From our own projects (our prices, not an industry benchmark):
- A focused web app replacing one or two core sheets (orders, stock or a party ledger): ₹90,000 to ₹2,00,000, usually 4 to 8 weeks, with your team trained in Hindi or English.
- Hosting: a small server for a business app typically costs a few thousand rupees a year.
- Maintenance: optional, about 15% of the build price a year for updates and support. Changes are quoted upfront.
Compare that with per-user tools over three years. At AppSheet's Core price of $10 per user per month, 15 users cost $1,800 a year, every year, and the data model stays inside Google's tool. A custom app costs more upfront and less later. Neither is always right; count your users and how long you'll use it.
Three famous spreadsheet failures
These are well-known cases, not statistics, but they show how ordinary the causes are:
- England, 2020: 15,841 positive Covid test results were missed in reporting, widely attributed to an old Excel file format hitting its limit (The Register).
- JPMorgan, 2012: the bank's own task force found a risk model run through spreadsheets with manual copy-and-paste and a formula that divided by a sum instead of an average.
- Reinhart and Rogoff, 2013: a much-cited economics paper's spreadsheet averaged a range that left out five countries (Herndon, Ash & Pollin).
If it can happen to a bank and a public health agency, it can happen to a stock register.
A checklist before you decide
- Count the people who edit the sheet each day.
- List the reports you build by hand and how long they take.
- Mark the formulas that encode business rules (special rates, schemes, credit terms).
- Check where the file lives and how it's backed up.
- Decide what must never be wrong: stock, dues, GST amounts.
If you'd like a second opinion, send us the sheet (with names removed) and we'll tell you honestly whether it needs software, a shared Sheet or nothing at all. See Excel to web app migration.
Common questions
Will we lose our existing data?
No. Data is imported from your sheets, checked against your totals, and you switch over only when the numbers match.
Can we still use Excel?
Yes. Any list or report can be exported to Excel for your CA or your own analysis.
Is Google Sheets enough?
Often, yes, for small teams. Shared Sheets with version history solve the "which file is latest" problem. They don't give you proper permissions, validation or a reliable audit log.
How long does migration take?
Most of our migrations take 4 to 8 weeks, including the parallel run.
How this guide was made: we used AI tools to help research and draft it. Every figure links to its source, and prices and limits were checked in October 2026. Prices change, so check before you buy.