Switching mortgage CRM without losing a deal
The fear that keeps brokerages on systems they dislike is not cost — it is the gap. The week where nobody quite knows where anything is and a live deal falls through it. Handled properly there is no gap, because you never stop working.
The one principle that prevents disasters
Never turn the old system off before the new one is verified. That sounds obvious and is violated constantly, usually because a contract end date drives the timeline rather than readiness.
Run both in parallel. Yes, it means brief duplication. That duplication is dramatically cheaper than discovering in week two that documents did not come across and the source system is already read-only.
Negotiate your old contract's end date after your migration is verified, not before. Vendors will often extend month-to-month rather than lose the renewal entirely — ask.
Step 1 — Get everything out
Export more than you think you need. Storage is free; realising six months later that call history did not come across is not.
- Contacts, including inactive ones and duplicates — sort them later, not now.
- Deals with their stage, dates, amounts, lender and product.
- Every custom field, even ones you believe are unused.
- Activity history — calls, emails, notes, timestamps.
- Documents, with their association to the right deal preserved.
- Tags, lists, segments and any automation rules you rely on.
If your current vendor makes exporting difficult, that is information about how the relationship ends — and a reason to start earlier than planned.
Step 2 — Map before you import
Mapping is where migrations are won or lost, and it is the step people rush. Your old system's 'Status' field and your new system's 'Stage' are rarely the same thing, and a bad mapping produces a pipeline that looks plausible and is wrong.
Do this on a sample first — fifty records covering your messiest cases, not fifty clean ones.
- 1
Map stages honestly
Old stages rarely map one-to-one. Decide deliberately where each lands rather than accepting a default, and write the decisions down — you will be asked why a deal is where it is.
- 2
Decide about dead data
Leads from 2019 that never responded do not need to come across as active. Import them archived, or not at all. Migrating rubbish makes the new system feel like the old one.
- 3
Handle duplicates before import
Merging is far easier on the way in than afterwards. The same person from two sources should arrive as one record with both histories attached.
- 4
Set renewal dates
If your old system did not track maturity dates, this is the moment to add them. It is the single highest-return field in a mortgage CRM and the migration is the cheapest time to populate it.
Step 3 — Verify against the source
Do not accept 'the import completed successfully'. Completing and being correct are different claims. Reconcile numbers you can check independently.
| Check | What to compare | Why it matters |
|---|---|---|
| Record counts | Contacts and deals in old vs new | Catches silent truncation on large imports |
| Pipeline value | Total active pipeline in both systems | Catches amount fields that mapped to the wrong place |
| Spot-check files | Ten real deals, field by field | Catches mapping errors counts will never reveal |
| Documents | A deal with many attachments | Documents are the most common thing to arrive detached |
| Activity history | A long-standing client's timeline | Confirms history came across, not just current state |
| Renewal dates | Deals maturing in the next year | Confirms the field you most need is actually populated |
Step 4 — Run in parallel, then cut over
Give it one to two weeks of overlap. New work goes into the new system; the old one stays readable. Agents keep a reference point, which removes almost all of the anxiety that makes migrations feel risky.
For a brokerage, migrate agent by agent rather than all at once. One agent live for a few days surfaces the problems that would otherwise hit forty people simultaneously.
- New deals start in the new system from an agreed date.
- In-flight deals finish where they started, unless they are early-stage.
- The old system becomes read-only rather than being switched off.
- One person is named as the go-to for the first fortnight.
- Cancel the old contract only after a full commission cycle has run cleanly.
The mistakes that actually cost people
- Cutting over on the first of the month, so the migration collides with commission processing.
- Migrating during your busiest season because the contract renewal happened to fall there.
- Importing every dead lead, making the new system feel as cluttered as the old.
- Skipping the sample import and discovering the mapping problem at full volume.
- Not exporting documents, then finding the old contract has ended.
- Training everyone in one session before anyone has real data to practise on.
The best week to migrate is the quietest one you can find with a full month before your next commission run. Pick the date around that, then negotiate the contract to fit.
Questions
How long does switching actually take?
A solo agent is usually live within a week. A forty-agent brokerage typically takes two to three weeks, and most of that is your team verifying data rather than the vendor moving it. The verification is the part worth not rushing.
Will we lose our client history?
Not if you export it and check it arrived. History is usually exportable; the failures come from not asking for it, or from accepting an import confirmation without reconciling against the source.
Should we migrate in-flight deals?
Early-stage ones yes, late-stage ones generally no. A deal about to fund has enough moving parts without changing the system it lives in. Let those finish where they started.
What if our current vendor will not give us our data?
Start with a written request citing your agreement, and check what your contract says about data ownership and export. In practice most vendors comply when asked formally. It is also a strong argument for confirming exit terms before signing with anyone new.
Do we need to train everyone before go-live?
Train briefly before, then again a week after with their real data in front of them. Pre-launch training on demo data is largely forgotten; the session that sticks is the one where people are looking at their own pipeline.
Send us whatever you have — a CRM export, a spreadsheet, a decade of Outlook contacts. You verify before anything goes live.
We do the migration
Send us whatever you have — a CRM export, a spreadsheet, a decade of Outlook contacts. You verify before anything goes live.