# Your Next Product Idea Is in Your Side Spreadsheet

Source: https://gravam.com/blog/your-next-product-idea-is-in-your-side-spreadsheet
Author: Tan Gravam
Published: 2026-11-15
Reviewed: 2026-11-15
Summary: The file you keep beside the official tool lists needs nobody wrote down. Read its column headers with AI, then check whether anyone else keeps the same one.

**You lie awake asking: what could I even build?** The textbook answer is to brainstorm, look for gaps in a market, pick a niche. It makes the idea sound like something out there that you have to find. For most people with a job, it's closer than that. **Most jobs with an official tool have a file kept beside it, because the tool doesn't do everything the work needs. Its column headers are a list of needs nobody wrote down.** You've been solving them by hand for years.

## The step: headers only

Open the spreadsheet you keep beside the official system: the tracker, the checklist, the "real" version of a report. Copy the header row. Only the header row, no data. Give it to your AI assistant:

```text
Here are my column headers. What does this tell you the official tool is missing?
```

You'll get back a list of needs. Read each one and ask: would someone in my role at another company keep this column too? It is the same move as splitting a weekly task into [what AI can do and what needs you](https://gravam.com/blog/what-to-learn-in-the-ai-era), applied to a file.

If you'd rather not paste anything into an AI tool, the free [Spreadsheet Column Audit](https://gravam.com/tools/spreadsheet-column-audit) does a rule-based version in your browser: paste the header row and it sorts the columns into missing fields, unwritten rules, judgement calls, copied data and notes. The headers you paste are not kept or sent anywhere.

## What the columns are telling you

Some kinds of column turn up again and again. I wrote them out in [The Spreadsheet Beside the System, Read Column by Column](https://gravam.com/blog/the-spreadsheet-beside-the-system); for a product idea, three matter most:

- **The corrected value.** "Real due date", "Actual amount". The system holds one truth and people work from another. Whoever closes that gap saves everyone a column.
- **The rule worked out by hand.** "Priority", "Tolerance", a colour code. If two people would fill it in the same way, it can be written down, and then it can be checked by software.
- **The decision recorded nowhere else.** "OK to pay?", "Approved by". The system doesn't know who decided, and someone needs to.

Each of these is a small product in the right market: a check, a template, a tool. Most of them should stay small.

## From a column to an idea

One sheet is one data point. An idea needs the need to repeat:

1. **Pick the column you'd most like to delete.** That's the most expensive one to keep.
2. **Ask three people in a similar role elsewhere** whether they keep the same thing. Not "would you buy this"; just "do you have a sheet like this, and what's in it?"
3. **If they do, write the smallest version first.** Often that's a guide or a template, the same first step as [turning the question you answer ten times into a guide](https://gravam.com/blog/turn-the-question-you-answer-into-a-guide).
4. **Decide how much you'll spend finding out.** A few evenings and a clear result that would tell you to stop. My own version of that is in [my kill criteria for product experiments](https://gravam.com/blog/my-kill-criteria-for-product-experiments).

## Where it goes wrong

- **Pasting the data.** Headers only. Leave out anything that names your company, clients or systems.
- **Building for your sheet alone.** If nobody else keeps the column, it's a fix for your team, not a product. That can still be worth doing, inside the company.
- **Automating the judgment.** An "OK to pay?" column needs a better record of the decision, not a robot that ticks it.
- **Starting with the app.** A template that removes one column is a better first test than software that replaces the whole sheet.

## What I do

I built the [Spreadsheet Column Audit](https://gravam.com/tools/spreadsheet-column-audit) from the same idea: the header row of the sheet beside an ERP or treasury system says more about what the system is missing than most requirement documents do.

Your idea isn't out there somewhere. It's in that file.

## Questions this article answers

**Q: How do I find a business idea from my job?**

Look at the workarounds you already maintain, starting with the spreadsheet you keep beside the official tool. Its column headers list what the tool is missing: a field, a rule nobody wrote down, a decision, data copied in by hand. If people in the same role at other companies keep the same columns, that need is not just yours, and it may be worth building for.

**Q: Is it safe to use my work spreadsheet to find a product idea?**

Use the column headers only, never the data, and leave out anything that names your company, clients or systems. The headers describe the kind of work, not your employer's information. Check your contract and company policy on side work and intellectual property before you build anything.

**Q: How do I test an idea I found in my own spreadsheet?**

Ask three people in a similar role at other companies whether they keep a sheet like yours and what is in it. If they do, you have a need that repeats. Then decide in advance how much time you are willing to put into a first version and what result would tell you to stop.
