Skip to content
NonStopExcel
← Articles

Why Is My Excel File So Slow? How to Fix "Calculating (12 Threads)" and Slow Workbooks

8 min read

You change one cell and the status bar says “Calculating (12 threads): 4%”. Then you wait. The file used to be fine, and now it freezes when you filter, takes a minute to save and makes everyone on the team nervous about touching it. When Excel is slow calculating threads, the message itself is not the problem. It is a symptom, and the cause is almost always something inside the workbook that you can find and fix.

This guide is the order I work through when a client sends me a slow file. Most of it you can do yourself in an afternoon.

What “Calculating (12 threads)” actually means

Excel splits recalculation across your processor’s cores. The number in brackets is how many threads it is using, which depends on your machine and on the setting under File > Options > Advanced > Formulas > Enable multi-threaded calculation. A laptop with more cores shows a higher number. The percentage is how far through the recalculation Excel has got.

So the message only tells you Excel is recalculating, and that the recalculation takes long enough for you to notice. Turning multi-threading off will not make the file faster. You need to find out why there is so much to recalculate.

Quick diagnosis: what exactly is slow?

Different symptoms point to different causes. Before changing anything, save a copy of the file and note which of these you’re seeing:

Symptom Most likely cause
Slow after every edit Heavy or volatile formulas
Slow to open or save, large file size Bloat: formatting, unused cells, objects, styles
Slow only when you click Refresh Power Query, connections or external links
Slow scrolling or filtering Conditional formatting, shapes, very large used range
Slow on one PC only Add-ins, 32-bit Excel, or the machine itself

Two quick tests narrow it down further.

First, switch to manual calculation with Formulas > Calculation Options > Manual. If editing becomes instant, the formulas are your problem. Press F9 to recalculate when you need to, and remember to switch back to Automatic afterwards, because a file left in manual mode shows stale numbers.

Second, press Ctrl+End on each sheet. Excel jumps to what it thinks is the last used cell. If your data ends at row 2,000 and Ctrl+End lands on row 900,000 or column XFD, the sheet is carrying a huge used range, usually from formatting that was applied to whole columns or rows at some point.

Remove bloat with Check Performance

In Microsoft 365, Excel can find and clean that formatting for you. Go to Review > Check Performance. The Workbook Performance pane shows how many cells are in use and how many can be optimized, sheet by sheet. Optimize sheet cleans one sheet; Optimize all cleans the whole workbook.

It removes formatting and similar data from cells that have no content. Microsoft warns that it will also remove deliberate formatting on empty cells, such as pixel art or a coloured layout with no values, so check the result and use Undo if something you needed disappears. Excel may also show a notification bar offering Check Performance when you open a workbook with a lot of unneeded formatting.

Check Performance does not catch everything. These are the other kinds of bloat I look for:

  • Conditional formatting rules that multiplied through copy and paste. Open Home > Conditional Formatting > Manage Rules, set Show formatting rules for to This Worksheet, and you may find hundreds of near-identical rules each covering a few cells. Delete them and recreate one rule per range.
  • Hidden shapes and objects. Home > Find & Select > Selection Pane lists every object on the sheet. Files that came from copied web pages or old templates sometimes hold thousands of invisible ones.
  • Old names. Formulas > Name Manager often shows hundreds of defined names, many with #REF! errors, inherited from copied sheets. Filter for names with errors and delete them.
  • Large pasted data that nobody uses any more. Old tabs with raw exports still recalculate and still get saved.

Saving as an Excel Binary Workbook (.xlsb) usually makes large files smaller and faster to open and save. It works with formulas, PivotTables and macros, but some third-party tools that read .xlsx directly cannot read .xlsb, so check before you switch a file other systems depend on.

Formula culprits

When manual calculation makes the file fast, the formulas need attention. Three patterns cause most of the trouble I see.

Volatile functions

Most formulas recalculate only when something they depend on changes. Volatile functions recalculate every time Excel recalculates anything, along with every formula that depends on them. Microsoft’s documentation lists NOW, TODAY, RAND, RANDBETWEEN, OFFSET and INDIRECT among them, and CELL and INFO depending on their arguments.

One TODAY() in a header cell is harmless. Thousands of OFFSET formulas building dynamic ranges are not. A typical example:

=SUM(OFFSET($B$2,0,0,COUNTA($B:$B)-1,1))

The same result without volatility, using INDEX to find the end of the range:

=SUM($B$2:INDEX($B:$B,COUNTA($B:$B)))

Better still, convert the data to an Excel table (Ctrl+T) and refer to the column by name. Tables grow automatically, so there is no need for a dynamic range at all:

=SUM(Sales[Amount])

INDIRECT is harder to replace because it is often used to pick a sheet by name. If you have many INDIRECT formulas pointing at monthly tabs, the usual fix is structural: put all months in one table with a Month column and use SUMIFS on it.

Whole-column references in array formulas

Functions such as SUMIFS and COUNTIFS handle a reference like A:A efficiently, because they only look at the used part of the column. SUMPRODUCT and other array calculations do not. This formula multiplies over a million rows every time it recalculates:

=SUMPRODUCT(($A:$A="Zagreb")*($C:$C))

It gives the same answer as this one, which is far lighter:

=SUMIFS($C:$C,$A:$A,"Zagreb")

Where you do need an array formula, point it at a table column or an exact range instead of the whole column.

Lookups repeated thousands of times

A sheet that pulls six columns from a 100,000-row table with six VLOOKUP formulas per row runs six full searches per row. Find the matching row once in a helper column, then reuse it:

=XMATCH($A2,Data[CustomerID])
=INDEX(Data[Region],$H2)

Here H2 holds the XMATCH result, and each INDEX simply fetches the value from that row, which costs almost nothing. In older versions without XMATCH, use MATCH($A2,Data[CustomerID],0) the same way. If the lookup table is sorted, XLOOKUP and XMATCH can use binary search (search mode 2), which is much faster again on large tables, but returns wrong results if the sort order is broken.

If the file is slow when it opens or refreshes, look outside the workbook.

External links to other workbooks force Excel to read those files, and links to files on a network drive or a path that no longer exists can hang the workbook for a long time. Go to Data > Edit Links (called Workbook Links in newer Microsoft 365 builds) to see every linked file. Remove links you don’t need, or use Break Links to turn them into fixed values. Breaking links cannot be undone, so do it on a copy first.

Power Query refreshes can be slow for different reasons:

  • Queries loaded to sheets that nobody reads. In Queries & Connections, right-click a query, choose Load To and set staging queries to Only Create Connection.
  • Many queries refreshing at once. Under each query’s Properties you can decide whether it refreshes with Refresh All.
  • Reading big source files over a slow network or a large SharePoint site. Filter early and remove unneeded columns in the first steps, so less data moves through the later ones.

Rebuild or patch?

Not every slow file should be patched. If the problems come from a few heavy formulas, a bloated used range or a pile of broken names, a cleanup usually solves it and leaves the structure people know.

A rebuild makes more sense when the slowness comes from the design itself: a tab per month linked with INDIRECT, raw data pasted into the same sheets as the calculations, or lookups stacked on lookups because the data never had a proper key. Speeding up such a file buys a few months, then it slows down again as the data grows. In those cases I move the data import into Power Query, keep calculations in tables and leave the report sheets for output only. The file usually ends up smaller and much easier for the next person to understand.

Frequently asked questions

Why does Excel say “Calculating (4 threads)” or “(12 threads)”?

The number is how many processor threads Excel uses for recalculation. It depends on your computer and on the multi-threaded calculation setting. The message means Excel is recalculating; the speed depends on what the workbook asks it to calculate.

How do I check what is making my Excel file slow?

Switch calculation to Manual to see whether formulas are the cause, press Ctrl+End to check for a huge used range, and run Review > Check Performance in Microsoft 365. Then look at conditional formatting rules, objects, defined names and external links.

Why is my Excel file slow to save?

Large files with heavy formatting, many objects or big pasted data take longer to write. Clean the bloat first, and consider saving as .xlsb. If calculation is set to Manual with Recalculate workbook before saving switched on, Excel also recalculates on every save.

Does 64-bit Excel help with large files?

It can, because 64-bit Excel can use more memory than 32-bit Excel. It will not fix slow formulas, though. You can check which version you have under File > Account > About Excel.

Is it safe to use Optimize all in Check Performance?

It is safe for normal data, but it removes formatting from empty cells, including formatting you may have added on purpose. Run it on a copy first and check the layout before replacing the original.

Send me the file

If you would rather not spend the afternoon on this, send me a copy of the workbook (an anonymised version is fine) with a short note on what is slow. I’ll look at it and come back with a fixed quote before any work starts. That is how my Spreadsheet Fix service works, from €350 for one workbook.