CRM build and migration, without losing the history
Clean records in, stages that match how your team sells, and every bit of history kept.
A CRM fails on dirty records and stages copied from a template, not the software. We clean and dedupe first, map stages to real buying decisions, move the history, and check adoption at thirty days.
Is it a fit?
Duplicates
The same customer under two spellings, which is why nobody argues with the report.
Stages
Each one matches something the buyer did, not something the seller hopes.
History
Old deals and notes move too, so the old system can be switched off.
Start here. Five answers on how information moves. You get back the week it costs and what that is a year.
Find the biggest leakWork it out in a minute
Five answers, returned in people rather than hours, and what that costs a year.
What you get
The records before the platform
Deduplicated, merged, and checked against a source you already trust. Two thousand contacts that are right beat nine thousand where a third are the same person. This is the step that gets skipped and the reason most launches lose the room in week two.
Stages that match the sale
Not a template. The points where a buyer decides something, named in the words your team already uses on calls. A stage nobody can tell they have reached is a stage nobody will update.
The history comes too
Old deals, notes, attachments and dates, mapped and moved, not left behind in a system you keep paying for so somebody can look things up. A migration that strands the past is not finished.
Adoption is the deliverable
Measured at thirty and ninety days: how many deals have a next step, how many were updated this week, how much still lives in somebody's inbox. A clean CRM nobody uses has cost you money and bought nothing.
Why the second CRM usually goes the same way as the first
A company on its second CRM almost always tells the same story. The first one was fine at the start, then the data drifted, then people kept a private spreadsheet because the CRM was wrong, and once two versions exist the official one loses. New software does not fix that. The same records go in, the same drift starts, and eighteen months later there is a third evaluation.
The drift starts with duplicates. The same customer under two spellings, a personal address and a work one, a deal logged by two people. Once a report can be wrong, nobody argues with the report, they just stop using it. So the cleaning comes first, before anything is chosen, and it is usually the least glamorous and most valuable week of the project.
Questions we get
Which CRM should we use?
Whichever one your team will update. For teams under fifty people the differences matter far less than the vendors need you to believe. It becomes a real question with unusual objects to model, heavy quoting, or records a regulator wants kept with the deal.
How long does a migration take?
A straightforward move runs four to six weeks, and about half of that is the data.
The variables are how many systems hold customer records today and how bad the duplication is. We size both in the first week so the date stops being a guess.
Will we lose anything?
Nothing that matters, and we tell you in advance what falls outside that.
Some things do not map cleanly: custom fields nobody has filled in for two years, attachments in formats the new system will not index. Those get listed and decided on, rather than discovered missing in month three.
Our team hates the current CRM. Will they use a new one?
Only if the reason they hate it gets fixed, and it is almost never the interface.
People stop updating a CRM when the effort has no return for them. If the pipeline view is wrong, the stages are meaningless, or they have to type something twice, they will keep their own list. That is the thing to fix, and moving platforms without fixing it just moves the problem.
On whether you need one at all, CRM or spreadsheet. On what the CRM should be able to see, business systems integration. A migration this size is project management as much as build.
If the customer record already lives in two places, what the second copy costs a year is usually the number that decides how urgent this is.
More in the guides and every answer in one place.
Read next
Whether you need one at all, what it has to be able to see, and where the numbers end up once it is right.
Shaheer leads the work, with engineers, writers, filers and analysts behind him. C-suite operations for a San Francisco AI company, Six Sigma on the process side, Anthropic certified on the Model Context Protocol, ten years across eight industries. See what we have built