SuiteProbe

Conversion

What survives a document conversion

Notes from our own conversion trials · 11 November 2026 · how we test

In short

Converting between document formats loses something every time. PDF to Word usually returns text boxes instead of flowing paragraphs, PDF to Excel adds stray rows around merged headers, and CSV to Excel silently strips leading zeros from postcodes and account numbers. Word to PDF is the only conversion that reliably preserves the original layout.

1. The four conversions people get wrong

Each of these is easy to run and easy to get subtly wrong. The pattern is the same every time: the conversion succeeds, the file opens, and something quiet has changed.

PDF → Word

What people expectEditable paragraphs that reflow like a normal document.

What usually happensMulti-column text usually comes back as separate text boxes rather than flowing paragraphs. It is editable, but it is not a Word document you would send to a client without rebuilding the layout.

PDF → Excel

What people expectA clean table with one value per cell.

What usually happensWorks well on ruled tables. Pages with merged headers or multi-line cells often gain empty rows, and numbers can arrive as text — which breaks every formula downstream.

CSV → Excel

What people expectThe data as it looked in the CSV.

What usually happensThe usual casualty is leading zeros: postcodes and account numbers become numbers, and the zeros disappear. Import as text, not as general, or you will lose them.

Word → PDF

What people expectExactly what was on screen.

What usually happensThe most reliable of the four. Fonts must be embedded, otherwise the file renders with substitutes on a machine that lacks them.

2. How to check a conversion before you rely on it

  1. Count the pages. If the output has more pages than the source, the layout moved.
  2. Search for a specific number you know. If a search for a value in the file returns nothing, it was converted to an image or lost.
  3. Click into a paragraph and press Enter. If the text is a box rather than a paragraph, editing will fight you.
  4. Check the first three rows of any imported table. Cells that should be text but became numbers are where the damage shows first.

3. The leading-zero problem, specifically

This deserves its own section because it is the most common silent data loss. A CSV containing 01234 opens in a spreadsheet as 1234. Nobody notices until a postcode fails validation or an account number does not match. The fix is at import time — set the column type to text before the data loads — not after, because the original value is already gone from the file in memory.

See the PDF tools included with WPS Office

Opens the official WPS site. Which tools are free and which need a paid plan is decided by WPS.

4. Task-specific guides

5. Sources