First in a series on the challenges of post-acquisition data migration.

Day 40 usually looks fine. Leadership is aligned, the branding refresh has shipped, and the all-hands meeting went well. Everyone feels good about where things stand.

Day 90 tells a different story. Finance is stitching two platforms together by hand. Sales is working out of two CRMs with no reliable way to tell which record is the real one. The migration team is still preparing for its first trial migration.

Not long after, most organizations arrive at what we call the Day 101 Problem: the moment it becomes clear that the Day 90 workarounds may quietly become the permanent architecture, and that the migration timeline everyone signed off on was fiction.

Hitting the Day 101 Problem is not proof that anyone scoped the project badly. It is simply an expectation management challenge. No one can know in advance how many trial migrations it will take before the data passes User Acceptance Testing (UAT), and no competent team signs off on cutover until the evidence says it is ready. Two dynamics guarantee this uncertainty: the onion problem, where every fix exposes a new layer of issues, and live system drift, where both companies’ systems keep generating new data between trial runs, so the target never holds still. No migration of real consequence wraps up by Day 101. Even the cleanest ones we have seen take months longer. Any go-live date set at kickoff is a guess at best, and sometimes off by an order of magnitude.

That is a hard message to deliver to a board that wants a single number for the budget, but the honest answer comes with caveats at every stage. Start with a rough range tied to known complexity: nine months for a smaller scope with clean data, twenty-four months or more for multiple enterprise-scale systems, eighteen months in between. Treat that range as a placeholder, not a commitment. After three or four trial migrations, usually somewhere around Day 60 to Day 100, the range narrows because you finally have real evidence of the order of magnitude involved. Narrower still is not the same as a hard number. You do not know you are ready until a trial migration clears UAT, whenever that turns out to be. The honest framing for a board is: the estimate gets sharper at each checkpoint, and it stays an estimate until it isn’t.

The one number that should never move is the UAT coverage threshold. Every trial migration gets measured against it, but only one trial ever clears it: the last one, and you only know that once it happens. That is what authorizes cutover. The date it happens is the one thing nobody can promise, not on Day 1 and not anywhere along the way.

More on this ahead: what you actually inherit when you close a deal, the most common pitfalls along the way, and what a realistic executive playbook looks like.

Where did you first realize your “temporary” workaround was not temporary anymore?

Originally posted on LinkedIn.