The Fork’s Confluence to Notion move is written up by the agency that did the work. Alegria.group’s case page says they moved more than 4,500 pages, about 10 GB, in two weeks.
I could not find a matching post from The Fork. The page names no Fork employee. It gives no project date. The numbers and the outcome lines below are Alegria’s.
The Fork is described there as the restaurant-booking app, a TripAdvisor subsidiary, in more than 20 European countries. The company-size field says more than 500 people.
If you are looking at the same route, start with the Confluence → Notion playbook. Read Alegria’s page as an agency report of one job, then check the losses the playbook says an import still leaves behind.
What they say the job was
Confluence had been used mainly by tech and engineering. The Fork wanted a wider knowledge base, so the rest of the company could use it too.
They also wanted the same information structure as the old platform, so the teams would recognize it. The 10 GB mixed text, images, links, usage rules, and videos. Alegria says that volume had to land in two weeks.
The three steps Alegria describes
They split the work into selection, import, and a sample check.
Before the move, they say they did not copy the space as it stood. They helped The Fork pick what was worth moving and restructure it.
For the move itself, they exported Confluence pages into a format Notion could take, then used RPA to automate the import.
After the import, they say they checked that the data was accurate and complete, using a manual sample. They also say they ran a QA review before the transfer and handled change management.
The page then says this produced a transparent migration and a perfect result. That sentence is the agency’s. The page does not show the sample size, a failed page, or a macro that broke.
What they claim changed
Alegria says the teams adopted Notion quickly, and that every team can use it as a base for how they work. They say information silos that did not talk to each other are gone.
They say The Fork now uses Notion for knowledge, both live and asynchronous, across those 20-plus countries. The claimed gains are collaboration, system documentation, governance, and decision processes.
One result line says user roles and page configuration were kept. The playbook says imported pages land private to the person who imported them. Confluence permissions stay behind. If Alegria preserved roles, that was a separate access pass after the import. The case page does not describe that pass.
What the page leaves out
There is no count of pages they refused to move, even though the pre-migration step was selection.
There is no list of macros, comments, or version history. There is no downtime note, because a documentation import often has none, and they do not say either way.
There is no cost. There is no quote from The Fork.
Mercari, on its own engineering blog, counted more than 50 import gaps and kept a Confluence archive. Alegria’s page has no matching failure section. This source reports the schedule and the outcome claims, and it leaves the broken pages out.
What I would copy
Keep the selection step. Alegria’s useful line is that they restructured before they imported. A two-week window on 4,500 pages is a reason to cut pages, not a reason to skip the sample.
Ask who checks the sample, and how big it is. “Complete and accurate” needs a number you can re-run.
Set permissions after the import. “Roles maintained” on this agency page still needs the access pass written down.
Use the Confluence → Notion playbook for macros, comments, history, and access. Use Alegria’s page for the shape of the job, and keep their outcome lines labeled as theirs.