Skip to content
DisrupsoftDisrupsoft
General

Enterprise CRM software: migrate without losing history

How to migrate enterprise CRM software to Dynamics 365 without losing history: what to clean up, how to map fields and keep your opportunities.

Enterprise CRM software: migrate without losing history

It's 9 a.m. on day one with the new CRM. A sales rep looks up her biggest customer and finds an almost empty record: the three opportunities lost last year are gone, so are the notes from the last meeting, and the created date says "yesterday". That customer has been buying for eight years. Everything was in the old system. In the new one, someone decided that "history can be migrated later".

Switching enterprise CRM software happens more often than people think: the company grows, or the system picked years ago no longer fits the process. The migration is where the sales team decides whether to trust the new tool. If reps can't find their history on day one, they go back to the spreadsheet. This guide lays out the work for a CRM migration to Dynamics 365, whether you're coming from Salesforce, HubSpot, a local system or a pile of spreadsheets: what to clean up, how to map fields, how to keep opportunities and activities, and how to run the cutover.

If you're starting from Excel alone, we covered that case in Customer management CRM: from Excel to a measurable process. This one is messier: an old CRM with years of data, custom fields and users who left long ago.

Before you move a single record, decide which history matters

The temptation is to "bring everything". In its best practices for migrating CRM data to Dataverse, Microsoft suggests a simple test: check how much each column is used in the source system and, if fewer than 5% of records have a value, ask the business whether it's worth keeping. The same guide says that in long-running CRM systems it's common to drop 30% to 40% of columns and up to 20% of tables.

Records work the same way. Some data matters but almost nobody opens it: emails older than three years, closed cases, lost opportunities, disqualified leads. You have three options for it:

OptionWhen it makes senseWhat to keep in mind
Migrate it to the new CRMThe team looks at it often, or it feeds trend reportsMore volume, more testing, longer load times
Leave it in the old systemIt's rarely consultedMicrosoft suggests keeping a few admin licenses and deactivating the rest; check what renewing them costs
Archive it elsewhere (a database or Azure Blob Storage)You need it for audits or contractsIt needs its own extraction process and someone who knows where to look

Our advice: migrate lost opportunities from recent years. They explain why a deal didn't close, and reps want to see them before quoting the same customer again.

Field mapping: the boring spreadsheet that saves you

For every source table, write down which Dynamics 365 table and column it goes to, what transformation it needs and who on the business side signs off. It isn't glamorous work, but it's where most load errors get prevented. A few issues show up in every project:

  • Value lists don't match. Stages, industries or loss reasons in the old CRM are rarely the same as in the new one. Microsoft recommends a lookup table that translates each source text into its target value instead of fixing records one by one.
  • Keep the source ID. Store the ID each record had in the old system in its own column. You'll use it to rebuild relationships during the load, to retry failed records and to answer "where did this customer end up?" months later.
  • Check text length. A notes field with no limit in the source may not fit the target column. Better to find out during mapping than halfway through the load.

How to keep opportunities and activities

This is where most migrations lose their history. The tool is rarely to blame. Usually it's the load order or the permissions.

Load order

Start with what depends on nothing (users, currencies, products), then accounts, contacts and opportunities, and leave activities for last, because their "regarding" field can point to any of those records. Microsoft's guide suggests sorting tables from fewest to most lookups. When two tables reference each other, insert them without that relationship first and fill it in on a second pass.

The original created date

If you do nothing, every record looks like it was created on migration day and your pipeline aging reports become useless. Dataverse lets you map the original date to the overriddencreatedon column: during import, that value becomes the record's created date. You need the prvOverrideCreatedOnCreatedBy privilege to do it. What you can't import is who created or modified the record, or the last modified date. If those matter to you, store them in custom columns.

Close opportunities at the end

Load won and lost opportunities as open and close them once the migration is done. According to Microsoft's guide, once closed they can't be changed without reopening them, and fixing a wrong amount gets slow.

Owners who no longer work there

What happens to the accounts of the rep who left two years ago? Dataverse lets you create unlicensed users (stub users) that can own records. The history keeps its original owner without paying a license for someone who's gone. Set the right owner when each record is created, because Dataverse security depends on that owner's business unit.

Attachments and documents

Contracts and proposals need a plan of their own. Microsoft suggests SharePoint for Office documents (up to about 20,000 files), Azure Blob Storage for larger volumes and a malware scan before moving anything.

Which tool to use, by size

ScenarioTypical tool
A few tables and clean CSV or Excel filesDynamics 365 import wizard
Data needs transforming as it loadsDataflows with Power Query
High volume, many relationships, repeatable loadsStaging database plus Azure Data Factory or another ETL tool

For complex migrations, Microsoft recommends a staging database. You take a snapshot of the source CRM, transform and validate it there, and only then load it into Dataverse. That way you can reconcile numbers before and after and rerun the load without touching production. One detail people forget: bulk loads are subject to Dataverse service protection limits, so your tool has to handle retries.

Testing and cutover

Cutover weekend shouldn't hold any surprises. To get there:

  1. Run a test load in a separate environment with an awkward sample: your largest customer, an account with dozens of contacts, opportunities in every stage.
  2. Ask three or four reps to look up their accounts. In an hour they'll find what no script catches.
  3. Reconcile numbers in source and target: account, contact and open opportunity counts, plus pipeline totals by stage.
  4. Do a full rehearsal and time it. That number decides your cutover window.
  5. From an agreed date and time, set the old system to read-only. Anything that changed after the rehearsal gets loaded as a delta.
  6. During the first week, have someone review load errors and team questions every day.

If the new CRM will also connect to your ERP, leave that integration until the migration is stable. We wrote about it in CRM for sales teams: how to integrate it with your ERP.

Migration and personal data

Chile's Law 21.719 is set to take effect on December 1, 2026. In October the Senate approved a bill that would postpone it, but by mid November it was still going through the Chamber of Deputies, so that's still the date. In practice: don't migrate personal fields nobody uses, decide who sees what in the new CRM, and document what was archived and where.

Frequently asked questions

Can I migrate the full history from another CRM to Dynamics 365?

Yes, including closed opportunities, activities and the original created date. What isn't kept natively is who created or modified each record; if you need it, store it in custom columns.

How long does it take to migrate enterprise CRM software?

It depends on volume, customizations and data quality. Cleanup and mapping usually take longer than the load itself.

Do I need licenses for reps who have left?

No. Dataverse supports unlicensed users that can own migrated records, so the history keeps its original owner.

Let's talk about your migration

At Disrupsoft we implement Dynamics 365 in stages: we start with a review of your sales processes, roll out module by module, connect it with Microsoft 365, Power Platform and your other systems, and support the team with training. If you're about to switch CRMs and don't want your customers to be "born" on cutover day, let's talk about Dynamics 365.

CrmDynamics 365Migracion De DatosVentas

← View more articles