Comparison · figures checked 1 October 2026
Test cases in a spreadsheet, and when to stop
Most test plans start in Excel or Google Sheets, and many should stay there. Here is where a spreadsheet is the right tool, the point where it starts costing you releases, and what changes if you move.
The facts about Google Sheets and Excel below come from Google's and Microsoft's own pages, read on 1 October 2026. The table links to each one.
| Tool | Free tier | Seats when free | The limit that bites | Who it suits |
|---|---|---|---|---|
| qarunbook | Yes | Up to 5 | Built for testing only, so less flexible than a sheet; it does not run automated tests. Paid plans are flat per workspace | Several people testing one plan on web, Android and iOS, with issues and retests tracked |
| Google Sheets | Yes, with a Google Account | Anyone you share with | No idea what a result, an issue or a retest is: every rule is a convention people have to keep | One or two people, one platform, a short plan, and everyone already in Google |
| Excel | No | Anyone you share with | Co-authoring needs Microsoft 365 and a file in OneDrive or SharePoint; otherwise copies multiply | Teams already in Microsoft 365, or a plan that has to be handed over as a file |
Why a spreadsheet works
Everyone already has one and knows how to use it. Google Sheets is available to anyone with a Google Account, several people can edit the same sheet at once, permissions control who can view, comment or edit, and revision history lets you go back to an earlier version. Excel co-authors the same way when the workbook lives in OneDrive or SharePoint and everyone has Microsoft 365. Add a column per platform, a dropdown for the result and conditional formatting, and you have a test plan by lunchtime. Sources: Google Sheets and Microsoft Support on co-authoring.
When you should stay
If one or two people test one platform from a plan of a few dozen rows, a spreadsheet is the right tool and a dedicated one is overhead. The same is true if the plan is a one-off, if the person who reads the results wants a file they can sort and filter themselves, or if the testing is really automated and the sheet is just a list of what the automation covers. Nothing here beats a spreadsheet for flexibility: you can add any column, any formula and any layout, and nobody has to learn a new tool.
Where it starts to break
- Platforms multiply. One result column becomes three for web, iOS and Android, then three more for the next build, and the plan gets copied to a new tab every release.
- Issues live somewhere else. A failed row says "see bug 214", the bug is in another sheet or a chat thread, and nobody is sure whether the fix was checked.
- A fix quietly becomes a pass. Someone changes Failed to Passed after a fix is deployed, without testing again, and the sheet cannot tell the difference.
- Who did what is hard to see. Google's version history shows changes over time, but seeing who changed a specific cell is limited to certain Workspace editions, and restoring an earlier version clears the cell-by-cell history before it. Source: Google Docs Editors Help.
- Progress is a formula someone maintains. How many checks have passed on iOS this build is a COUNTIFS that breaks the first time a column moves.
What qarunbook is
A workspace for manual and acceptance testing, free to start, that keeps the parts of a spreadsheet that work. You import the sheet you already have, as Excel or CSV, or a markdown plan, and it becomes a board of checks across the platforms you named. Each check records its result per platform, with the name of whoever recorded it. Issues sit on the check that found them, and marking an issue fixed sends the check back for a retest instead of letting it pass on its own. Progress is counted for you. When you need a file again, a report exports to PDF, Excel or CSV.
Once an app is live it can run as bug tracking instead: issues without a check, with a priority, an assignee and a fix that someone confirms. It ships a remote MCP server, so Claude Code or Cursor can read what is failing, work the checks it can drive and record results under its own name; see the QA MCP server guide. Pricing is flat per workspace, never per seat, with a free plan for two apps and five teammates, and a REST API from the Team plan up. Source: qarunbook pricing.
What qarunbook is not
- As flexible as a spreadsheet. A check has steps, an expected result and a result per platform. If you need arbitrary columns and formulas, a sheet is better at that.
- A test automation tool. It does not run automated tests and has no JUnit or XML ingestion.
- Free for any number of people. The free plan holds five teammates. A shared sheet holds as many as you share it with.
- Offline. It is a hosted web app. Google Sheets can be edited offline; qarunbook cannot.
What moving actually involves
Very little. Add an app, give it the spreadsheet you already have, and AI reads it into sections and checks, whatever its columns, and you review every check before it is saved. There is also an Excel and CSV template on the import screen if you would rather match a layout. Old results stay in the sheet, which is where they were useful anyway. If you are starting a plan from nothing, the UAT test plan template is a reasonable first draft, and the free tools comparison covers the dedicated tools.