From Excel and WhatsApp to a Simple Web App: What to Build First
By Kumar Deepansh ·
Move to a simple web app when the same data lives in several spreadsheets and chat threads, and people keep asking each other “which version is right?”. Start with the one workflow that costs you the most time or money each week, usually stock, quotes or follow-up reminders, and build only that. Keep Excel and WhatsApp for everything else until the first piece is working.
Why do Excel and WhatsApp stop working?
They do not stop working all at once. Each is good at what it was made for: Excel for calculation, WhatsApp for talking to people. The trouble starts when they become your database and your workflow engine.
- No single source of truth. Stock counts live in one file, prices in another, and the real status of an order lives in a chat thread.
- No rules. A spreadsheet will accept a negative quantity, a typo in a customer name, or a deleted formula without complaint.
- No history. When a number changes, you cannot see who changed it or what it was before.
- Everything depends on one person’s memory. Follow-ups happen if someone remembers.
On the error side there is real research. Raymond Panko, who has spent years studying spreadsheet errors, summarises field audits of real organizational spreadsheets: older audits found errors in 24% of 367 spreadsheets, but the more recent audits, with better methods, found errors in at least 86%. He also notes that most audits only reported substantive errors, so the true figure for large spreadsheets is probably not lower. Those are audits of larger working spreadsheets, not a prediction for yours, but they show why “it works fine” is not proof it is correct.
What does a small business actually need first?
In the small businesses I look at, the first useful app is usually one of three things. Pick one, not all three.
| Need | What it replaces | Smallest useful version |
|---|---|---|
| Inventory | A stock sheet that is out of date by Wednesday | Items, quantity on hand, one screen to add or remove stock, a low-stock flag |
| Quotes | Word or Excel quotes copied from the last customer | Customer list, price list, quote built from line items, a PDF to send, a status (sent, accepted, lost) |
| Reminders | “Remember to follow up” in a chat | A list of follow-ups with due dates and an owner, plus a daily list of what is due today |
If you cannot decide, ask which one has cost you money in the last three months. A missed follow-up that lost a sale, or an oversold item, is a better guide than a feature wish list.
How do I decide the scope?
Write the scope on one page before anyone writes code. I use these steps:
- Name the workflow, not the software. “From enquiry to accepted quote” is a scope. “A CRM” is not.
- List the people who touch it. Owner, sales, warehouse, accountant. Each one needs a way in, and each one needs only their part.
- List the records. Customers, items, quotes, stock movements. If a thing is not a record someone needs to look up later, it probably does not belong.
- Copy the real data. Look at the actual spreadsheet and the actual chat history. The columns people really use are the requirements.
- Set a “done” test. For example: “I can create a quote in under two minutes and see its status without asking anyone.”
Then freeze it. New ideas go on a “later” list. That list is where most scope creep gets caught before it costs money.
What should I build first?
Build the thing that gives you one trustworthy list. For inventory that is the stock count. For quotes it is the customer and price list. For reminders it is the follow-up list.
A good first release has:
- Login with at most two or three roles.
- One main table with search and filters.
- Create, edit and a clear history of changes.
- An import from your current spreadsheet, so day one starts with real data.
- An export to CSV, so you are never locked in.
Run it next to the spreadsheet for a short time, compare the two, then retire the spreadsheet for that workflow. Retiring the old way matters. Two sources of truth is worse than one messy one.
What about WhatsApp?
WhatsApp is where your customers already are, so do not try to replace it. The better move is to keep the conversation there and keep the record in the app.
If you later want the app to send messages, check the rules first. According to Meta’s developer documentation, businesses on the WhatsApp Business Platform are charged per delivered message, with the rate depending on the message category and the recipient’s country. Template messages fall into marketing, utility and authentication. A 24-hour customer service window starts when a customer messages you, and outside that window you can only send pre-approved template messages. Pricing pages change, so read the current one before you plan costs. As of November 2026, the documentation I read shows pricing updates effective 1 October 2026, including charges for service messages beyond 1,000 per month per business phone number.
For a first version, a link that opens a WhatsApp chat with a prefilled message is often enough, and it needs no messaging platform.
What should I not build?
- Anything you do only twice a year. A spreadsheet is fine for it.
- Custom reports for every question. Start with a table you can filter and export.
- A mobile app. A web app that works on a phone browser covers most small teams.
- Full accounting. Keep your accounting software and send it clean data.
- Automation of a process nobody has written down. If the team cannot describe the steps, a program cannot follow them.
- Every integration at once. Add one at a time, after the core is trusted.
FAQ
Is a web app better than a no-code tool or a ready-made SaaS?
Not always. If a ready-made tool fits your workflow with few changes, use it. A custom app makes sense when your process is specific enough that you keep working around the tool.
How long does a first version take?
It depends on the scope, so I price a fixed scope after reading your current spreadsheet. A single-workflow first release is a much smaller job than a full system, which is why I suggest starting there.
Can I keep using Excel for some things?
Yes. Keep it for ad hoc analysis. Use the app for the records that other people depend on.
What happens to my existing data?
It gets imported on day one, and the app should always let you export everything back to CSV.