What a migration audit actually finds

A scoped migration audit is a read-only pass over an export from the system you already run. It produces five things, and four of them are boring.

The five outputs

Row counts per record type. Dedupe estimate. Field map. Data-quality findings. A fixed migration date.

Row counts tell you how big the job is. The field map tells you what has nowhere to land. The date tells you when it can happen. Those three are logistics and they matter, but they do not change how anybody feels about switching.

The one that does

The dedupe estimate is different, because it is the first honest look most principals ever get at their own database.

Here is the arithmetic, and you can run it on your own export without us. Take your contacts table. Count rows. Now count distinct mobile numbers, ignoring blanks and normalising to a single format. Then count distinct email addresses the same way. Then count rows where both are blank.

Four numbers. On the exports we have looked at, the pattern repeats: total rows sits well above distinct contactable identities, and the gap is not evenly spread. It concentrates in the records created during a portal enquiry burst, because an enquiry with a slightly different spelling of a name creates a new record and nothing ever merges it back.

Why the gap matters more than its size

Your database size is a number you quote. Your contactable identity count is the number your email tool bills against, the number your nurture campaign actually reaches, and the number your appraisal pipeline is really drawn from.

If those two numbers are far apart, you are paying for one and running on the other. That is not a vendor problem and switching does not fix it by itself. What switching does is force the question, because a migration that carries duplicates forward has moved the problem rather than solved it.

What we do with it

We show you the estimate and the sample that produced it, before you commit to anything. You can disagree with any merge and we will carry both rows.

We do not charge for the audit and we do not require anything in return for it. That is not generosity. It is the cheapest way for us to find out whether your data is a job we can do well, and the cheapest way for you to find out whether we can read your export at all before you have signed anything.

Reproduce this

Every figure in this piece comes from the arithmetic above, run on your own export. We have published the method rather than a headline percentage, because a percentage from someone else's database tells you nothing about yours.