Every migration we look back on with regret started the same way: a tool was selected, and only afterwards did anyone ask what the content actually was. The selection meeting is enjoyable. It has demos, comparison grids and a decision at the end of it. The alternative first step, which is reading six hundred pages of your own material, has none of those things.
But the tool is almost never the constraint. The constraint is that nobody owns the entry requirements field, that three teams have been publishing contradictory versions of the same policy for four years, and that the review date on most pages is a polite fiction. Move all of that into a new system and you have paid a large sum to reproduce the problem somewhere more expensive.
We now insist on governance before templates. One owner per field, an approval route, a review date that somebody actually keeps. It is a dull month and it makes the migration itself almost boring, which is the entire point. The projects that go badly are the ones where the interesting decisions were taken first and the boring ones were left until after launch.
Comments