Microsoft's guidance: convert your files before 1 October 2026. Microsoft's own published date, given here as honest context — not a countdown to a sale.

Small organisations have a particular version of this problem, and most advice on the internet is not written for it. You do not have an IT department. You may not have a budget line for this at all. The person who made the newsletters for fifteen years may have retired, and their PC may have gone with them. And the files themselves are not just documents — for a lot of churches and schools the Publisher folder is the institutional memory.

So here is a plan sized for that reality.

Afternoon one: find out how big this actually is

Do not start converting anything. Start by counting, because the number determines everything after it and people are routinely wrong about it by a factor of five.

On a Windows PC, open PowerShell and run:

Get-ChildItem -Path "C:\" -Filter *.pub -Recurse -File -ErrorAction SilentlyContinue | Measure-Object

Do it again for each network drive and each old external hard disk in the cupboard, and write the actual number down. Under about thirty files, you can do the whole job by hand this afternoon and skip most of the rest of this page. Save a list of the files too, because you will want it to check your work against at the end.

The question that changes everything: is Publisher still installed anywhere?

Before choosing any method, walk round the building and find out. Check the office PC, the vestry laptop, the machine in the school library, the one in the finance cupboard that nobody has switched on since 2021.

  • If yes — that machine is your most valuable asset for this job and it will not be there forever. Microsoft's free Convert-PubFileToPDF.ps1 script can drive it over your whole archive unattended; we walk through it in the bulk conversion guide. Do this before the machine is replaced.
  • If no — Microsoft's script cannot help you, because it works by driving Publisher rather than reading the files itself. You need a different route, covered further down.

Afternoon two: split the pile in two

This is the step that turns an impossible job into a finite one, and almost everybody skips it.

Pile A — archive. Things you need to be able to read again: eleven years of newsletters, old service sheets, past prospectuses, the 2015 fundraising appeal. Nobody is ever going to edit these. They need to be readable in 2040, nothing more.

Pile B — live templates. Things you make a new version of every week or term: the weekly bulletin, the newsletter masthead, the gift-aid form, the letterhead. These need rebuilding in whatever you are moving to.

Nearly every organisation that does this discovers Pile A has hundreds of files and Pile B has four. That is the good news, and it is why sorting first matters: if you try to rebuild all 900, you will stop in week two and nothing will get done.

While you sort, do the governance bit. Some of those old files contain personal data — membership lists with home addresses, pupil names, donor records, safeguarding contacts, rotas with phone numbers. For those, the right question is not only "how do we convert this" but "should we still be holding this at all". If your retention policy says no, deliberate deletion is the correct action and it saves you converting them. Doing this now, while you have the files in front of you, is far easier than doing it later — and it is a genuinely good outcome from an otherwise annoying deadline.

It also settles the online-converter question for you. Anything in that category should not be uploaded to a third-party server, however convenient.

Afternoon three: convert Pile A

For the archive pile, PDF is the right target. It preserves the appearance of the page, opens anywhere, needs no special software, and will still open in twenty years.

Three routes, in order of preference:

  1. Publisher's own File → Save As → PDF, driven in bulk by Microsoft's free script if you have a Publisher install. Best fidelity, free, unattended.
  2. The free LibreOffice Draw, if you do not. Try it on five representative files and look at the results honestly before committing the archive to it — simple documents usually come through well, complex ones vary.
  3. A local recovery tool, if neither of those fits. See below.

Whichever you use, work from copies and never delete the original .pub files. Copy the whole folder to an external drive first. Support for old formats tends to improve over time rather than disappear, and the original is the only complete copy of the document in existence.

Then reconcile. Count the PDFs and compare against the inventory from afternoon one. A bulk conversion that skipped forty files while reporting no obvious problem is the classic way to believe a job is finished when it is not.

The bit people regret skipping: get the ingredients out too

A PDF is a photograph of the page. It is not the material the page was made from.

In two years somebody will want the logo for the new website, or the photographs from the 2009 centenary, or the exact wording of a standing policy that was printed in the back of a handbook. Pulling those out of a PDF is fiddly and lossy: text arrives with line-break hyphens in the middle of words, and images come back re-compressed at whatever resolution the export used.

So alongside each PDF, keep the text and the images as separate files. It is a small amount of extra effort at conversion time and it is the difference between an archive you can read and an archive you can use.

Old logos are the standout case. In Publisher-era documents these are very often WMF or EMF metafiles — vector images, which means a properly extracted one can be scaled up to a banner without going blurry. A version screen-grabbed from a PDF cannot. If your organisation still uses a crest or logo that was drawn in the 1990s, recovering the vector original may be the single most valuable thing to come out of this whole exercise.

One caveat worth having before you plan around it. Our own tool decompresses and writes out WMF and EMF metafiles, but across the six genuine Publisher files it has been run against, 20 image files came out — 15 PNGs and 5 BMPs rebuilt from pattern-fill bitmaps — and not one metafile, because none of those documents contained one. Test the route on one of your own files that you know has a vector logo in it before committing the archive to any particular method.

If Publisher has already gone from the building

This is the situation that the retirement will create for a great many small organisations, and it is what we built our tool for.

Publisher Rescue reads your own .pub files directly on your own computer, with no copy of Publisher needed and nothing uploaded anywhere. From each file it recovers your text as a Word .docx plus RTF and plain text; the pictures stored in the file as standard image files, with PNG, JPEG and TIFF written out as stored, Windows bitmaps rebuilt as openable BMPs, and WMF and EMF metafiles decompressed and exported properly; a PDF of the recovered content ending in a notes page listing what could not be rebuilt; the page size and the number of pages actually rebuilt, which is the number of layout pages in that PDF rather than a claim about the original; the Publisher format family the file was saved in; and a report — in plain language and in JSON — of exactly what was and was not recovered.

Honest scope, because you should not buy the wrong thing. It recovers content, not page design. The .docx is text-faithful — your words come through reliably — but it is not a replica of the original page, and the words arrive as a plain flow: across the six genuine Publisher files this build has been run against, no character or paragraph format table could be decoded, so sizes, bold, italic and colour are not carried over and no font name was reported. The PDF is not a reproduction of your design either — fonts, sizes, bold/italic and colours are not reproduced, and WordArt, master-page artwork, colour schemes, shape fills and table cell layout are not rebuilt; the PDF states that on every page and lists what it left out on its own notes page. If what you need is a faithful-looking page, a PDF exported from Publisher itself is the better output, and if you can still get one you should. Coverage is Publisher 2000–2019, and the tool names the format family a file was saved in rather than the exact release — Publisher 2002 through 2019 all write the same version marker. The Publisher 2000 path is in the build but no genuine Publisher 2000 file has been through it, and its older page-layout directory is not read; Publisher 97/98 and earlier are a different format and are reported as out of scope rather than mis-converted.

Your original is only ever read, never written to: the report records a SHA-256 checksum of the source taken before the scan and again afterwards, both printed in full so you can compare them yourself.

Supported operating systems: Windows 10 and 11 (64-bit), and macOS 10.15 or later on Intel and Apple silicon Macs. One purchase covers both — you get each download.

Two things about it that matter specifically for this audience: it runs on your own machine with no account — your files never leave it; the app goes online only to check your licence key, and works offline for up to 21 days between checks, so member and pupil data never leaves your premises; and it is a one-time purchase rather than a subscription, which suits an organisation doing a one-time job on a one-time budget. Saving the recovered files is what the £79 licence unlocks.

Afternoon four: rebuild Pile B, and write it down

Now rebuild your handful of live templates in whatever you are moving to — Word, PowerPoint, Canva, Scribus, Affinity Publisher, whatever your volunteers can actually use. Choose on that last criterion above all others: the best tool is the one the next person can pick up.

Then write a single page and put it in the shared folder: where the originals are, where the PDFs are, where the extracted text and images are, what the new template is made in, and who to ask. Half of what makes an archive useless in ten years is not the file format — it is that nobody remembers where anything went.

The whole plan on one line each

  1. Count the files. Write the number down.
  2. Find out whether any machine still has Publisher. This decides your method.
  3. Split into archive files and live templates.
  4. While sorting, decide what you should not still be holding, and delete it deliberately.
  5. Copy everything to an external drive. Work only from copies.
  6. PDF the archive pile, then reconcile the counts.
  7. Extract the text and images as well — especially old vector logos.
  8. Rebuild only the live templates.
  9. Write the one-page note. Keep the originals forever.