Why is my Excel file so large?

Rarely the data. The usual causes are a used range that extends to row 1,048,576 because of stray formatting, conditional formatting applied to entire columns, pivot table caches storing a second copy of the source data, images pasted at full camera resolution, and unused named ranges or defined styles accumulated over years of edits. A spreadsheet with 100,000 rows of genuine numbers is typically only a few megabytes.

How do I compress an Excel file to a smaller size?

Work in this order: reset the used range on every sheet with Ctrl+Shift+End to find it and delete everything beyond your data; remove conditional formatting applied to whole columns; clear pivot caches by unticking 'Save source data with file'; compress embedded images through Picture Format; then save as .xlsb. The first two steps alone often cut the file by 80% or more.

Does saving as XLSB make an Excel file smaller?

Yes, usually by 30 to 70%. XLSB stores the workbook as a binary format rather than compressed XML, so it is both smaller and noticeably faster to open on large files. The tradeoffs are that some external tools and data pipelines cannot read .xlsb, and macro-blocking policies sometimes treat it with suspicion because it can carry macros.

Should I zip an Excel file to email it?

It is worth trying, because XLSX is XML underneath and spreadsheet data compresses well — a ZIP often takes another 10 to 20% off. But XLSX is already a ZIP archive internally, so the gain is much smaller than people expect. Fixing the used range or the pivot cache will save far more than zipping the result.

Why is my Excel file large when it has almost no data?

Almost always a phantom used range. Selecting entire columns to apply formatting, borders or conditional formatting makes Excel treat over a million rows per column as in use, and it stores a cell record for each. Press Ctrl+Shift+End — if the cursor lands thousands of rows below your actual data, that is the whole problem.

How to Compress an Excel File (and Why XLSX Files Get So Large)

By Hieu DinhPublished August 22, 2026

How to Compress an Excel File

Numbers are cheap. A worksheet with 100,000 rows of genuine data is a few megabytes, because .xlsx is XML under ZIP compression and numeric data compresses extraordinarily well.

So when a workbook is 80 MB and contains a few thousand rows, the data is not responsible. Something else is inflating it, and the fastest way to compress an Excel file is to identify what before changing any settings — because the fixes are not interchangeable. Compressing images achieves nothing if your problem is a phantom used range.

Look Inside the File First

Like .docx and .pptx, an .xlsx file is a ZIP archive. You can open it and see exactly where the weight is.

Duplicate the workbook. Work on the copy.
Rename it from budget.xlsx to budget.zip.
Expand it.
Look at xl/worksheets/ and xl/media/.

What you find tells you which section below to read:

What is large	Cause	Fix
One `sheetN.xml` at 40 MB+	Phantom used range or column-wide formatting	Causes 1 and 2
`xl/media/` full of photos	Images at full resolution	Cause 4
`pivotCache` files are enormous	Pivot tables storing duplicate data	Cause 3
`styles.xml` unusually large	Accumulated junk styles	Cause 5

That five-minute check saves you from applying four fixes when one was needed.

Cause 1: The Phantom Used Range

This is the single most common reason an Excel file is large, and the most disproportionate — it routinely turns a 400 KB workbook into 60 MB.

Excel tracks a used range per sheet: the rectangle from A1 to the furthest cell it believes contains something. Select column D and apply a border, and Excel now considers all 1,048,576 rows of that column to be in use. It writes a cell record for every one of them.

Diagnosing it

Open a sheet and press Ctrl+Shift+End. The cursor jumps to the last cell of the used range.

If your data ends at row 3,000 and the cursor lands on row 900,000 — or in column XFD — you have found your file size.

Fixing it

Click the row number immediately below your last row of real data.
Press Ctrl+Shift+↓ to select every row to the bottom.
Right-click → Delete (delete rows, not just clear contents).
Repeat horizontally: click the column right of your data, Ctrl+Shift+→, delete columns.
Save, close, and reopen the file. The used range is only recalculated on save.

That last step matters. Skip it and Ctrl+Shift+End still reports the old range, and the file appears unchanged.

Repeat per sheet. On a workbook with fifteen sheets built by different people, several usually have this problem.

Cause 2: Conditional Formatting Applied to Whole Columns

The close cousin of the phantom range, and it is worse because conditional formatting rules are stored per applied range and re-evaluated constantly. A workbook with column-wide rules is both large and painfully slow.

Home → Conditional Formatting → Manage Rules → Show formatting rules for: This Worksheet.

Look at the Applies to column. Anything reading $A:$A or $A$1:$A$1048576 should be narrowed to the actual data range, $A$2:$A$3000.

You will also frequently find the same rule duplicated dozens of times over adjacent ranges — the residue of copying and pasting formatted rows. Delete the duplicates and apply one rule across the whole range.

Cause 3: Pivot Table Caches

Every pivot table stores a cache: a complete second copy of its source data, embedded in the file. A pivot over 200,000 rows carries those 200,000 rows twice.

If the source data lives in the same workbook, you are storing it twice for no benefit.

Right-click the pivot table → PivotTable Options.
Data tab.
Untick Save source data with file.
Tick Refresh data when opening the file.

The pivot rebuilds from the source on open. Do this for every pivot whose source is in the same workbook. On finance and reporting workbooks this is often the largest single saving available.

Cause 4: Images at Full Camera Resolution

Excel stores what you give it. A logo dropped in from a 24-megapixel photo stays 24 megapixels no matter how small the displayed image is.

Click any image.
Picture Format tab → Compress Pictures.
Uncheck "Apply only to this picture".
Tick Delete cropped areas of pictures.
Choose a resolution by destination:

Destination	Resolution
On-screen only	96–150 ppi
Printed reports	220 ppi
Never	"High fidelity"

Excel's compression is a blunt instrument — it applies one setting to photographs and screenshots alike, and screenshots degrade badly under it. If the workbook's images matter, compress them properly first and insert the optimised versions.

Cause 5: Accumulated Junk

Long-lived workbooks collect debris that nobody ever removes.

Unused styles. Every copy-paste between workbooks imports style definitions. Files that have circulated for years accumulate tens of thousands. A large styles.xml in the unpacked archive is the tell.

Defined names. Formulas → Name Manager. Delete anything showing #REF! — those are broken references to deleted sheets, and they persist indefinitely.

Hidden sheets. Right-click any tab → Unhide. Old workbooks routinely hide a 200,000-row raw data sheet nobody has needed since 2019.

Unused formatting on empty rows. Covered by the used-range fix above, but worth re-checking after deleting sheets.

Cause 6: Volatile Formulas and Full-Column References

=VLOOKUP(A2, Sheet2!A:Z, 5, FALSE) scans a million rows. Multiply by 10,000 formulas and you have a workbook that is both bloated with calculation chains and slow enough to be unusable.

Replace full-column references (A:Z) with bounded ranges ($A$2:$Z$5000) or structured table references. The file-size gain is modest; the performance gain is not.

Where a lookup result never needs to change again — last quarter's closed figures — copy and Paste Special → Values. Formula records vanish, and the numbers remain.

Should You Save as XLSB?

The Excel Binary Workbook format stores the workbook as binary rather than compressed XML. It is typically 30–70% smaller and opens noticeably faster on large files.

File → Save As → Excel Binary Workbook (*.xlsb).

Worth knowing before you commit:

Many external tools, Python and R libraries, and BI connectors read .xlsx but not .xlsb.
It can contain macros, so some corporate security policies treat it with suspicion.
It is not a substitute for the fixes above. Saving a workbook with a phantom used range as .xlsb gives you a smaller file that is still structurally broken and still slow.

Fix the causes first. Then decide whether .xlsb is worth the compatibility cost.

About Online Excel Compressors

Plenty of web tools will compress an XLSX. They unpack the archive, re-encode the embedded media, strip some metadata, and repack it — a subset of the operations above, run on someone else's server.

They cannot fix a phantom used range, a pivot cache, or column-wide conditional formatting, which is where the size usually is.

And there is the question of what Excel files contain. Payroll, customer lists, financial models, pricing, unreleased results. Uploading those to a free service means handing them to a third party under a retention policy you have not read. For anything confidential — which describes most serious spreadsheets — do the work locally.

A Practical Order of Operations

Duplicate the file.
Unpack the ZIP to find out what is actually large.
Reset the used range on every sheet, then save, close and reopen.
Narrow conditional formatting to real data ranges.
Clear pivot caches where the source is in the same workbook.
Compress or replace images.
Delete junk — broken names, hidden sheets, unused styles.
Save as .xlsb if compatibility allows.

Steps 3 and 5 do the overwhelming majority of the work on most real workbooks.

Fix the Container, Then the Contents

The pattern is the same one that governs Word documents and PowerPoint decks: Office files are ZIP containers, and what makes them large is almost never the thing you were actually working on.

For spreadsheets specifically, the bloat is usually structural — ranges, caches, formatting — and free to remove. What is left after that is embedded media, and media responds to real compression.

Compresto compresses the images, video and PDFs that go into and come out of your workbooks, natively on macOS — batch processing for a folder of charts and figures, format-appropriate handling for screenshots versus photographs, and no uploads, which matters when the spreadsheet is confidential.

Download Compresto for Mac →