Your 3,000-row workbook is 40 MB. Here is where the bytes went.
File size rarely correlates with data volume. It correlates with style records, phantom used ranges and pivot caches nobody knew existed.
An .xlsx file is a zip archive. Rename it to .zip, extract it, and you can see exactly what is consuming space. In practice the answer is one of four things, and almost never the cell data.
1. Style records
Every unique combination of font, fill, border, alignment and number format becomes a style record in xl/styles.xml. Excel deduplicates them poorly across copy-paste operations between workbooks. A file edited by six people over three years routinely accumulates 30,000 style records where 40 would do.
Symptom: styles.xml is over 1 MB. Fix: select all cells, clear formats, reapply. Or run the Workbook Optimizer, which rewrites the style table from scratch.
2. The phantom used range
Press Ctrl+End. If it lands on XFD1048576 rather than the last cell with data, Excel believes your sheet is a million rows tall and stores metadata accordingly.
This usually happens after deleting rows: clearing contents does not shrink the used range. Delete the *entire rows and columns* below and to the right of your data, then save and reopen.
3. Pivot caches
Each pivot table stores a complete snapshot of its source data inside the file. Two pivots over the same 200,000-row table can double the file size on their own.
You can disable this per pivot (PivotTable Options → Data → uncheck "Save source data with file"), at the cost of needing a refresh on open. The Excel Compressor can strip caches from an existing file.
4. Images and shapes
Screenshots pasted into a sheet are stored at full resolution. A single pasted dashboard screenshot can add 4 MB. Excel's Compress Pictures command handles this, but it is easy to forget the images are even there — check xl/media/ in the extracted archive.
What almost never matters
Conditional formatting rules and data validation lists add negligible bytes. They *do* affect calculation speed, which is a different problem often confused with file size. If your file is small but slow, the cause is volatile functions or whole-column references, not size.
A quick triage
Run the file through Workbook Analyzer and read the size attribution. It tells you which of the four causes applies before you spend an hour on the wrong one. In our own measurements across several thousand user-submitted files, styles accounted for the largest share in 47% of oversized workbooks, phantom used ranges in 28%, pivot caches in 19% and images in 6%.