Moving from spreadsheets to Salesforce without losing your mind
Data migration is where most rollouts get painful. A plain checklist for cleaning leads, accounts, and deals before you hit import.
Nobody wakes up excited about a Salesforce data migration. They want the CRM working. But if you import dirty spreadsheets on a Friday and go live Monday, you will spend the next six months apologizing for duplicate accounts.
We have done migrations from HubSpot, Zoho, Pipedrive, and yes — a lot of Excel tabs that lived on someone's desktop. The pattern is always the same: the data looks fine until you try to match it to Salesforce objects.
Clean before you map
Do not start in Salesforce. Start in the spreadsheet.
- Pick one source of truth per company. If "Acme Ltd" and "Acme Limited" are the same customer, merge them now — not after import.
- Decide what an Account vs Contact vs Lead means for your business. Founders often mix them because the spreadsheet did not care. Salesforce does.
- Kill columns you will never use. That "notes from 2019 trade show" column can live in an archive file. Reps ignore forms with forty fields.
Mapping is a conversation, not a template
Salesforce gives you standard objects for a reason. Fighting them costs time.
Ask: "When a deal closes, what do we need to report on?" That answer drives Opportunity stages, not the other way around. If your old CRM had "Proposal sent" and Salesforce has "Negotiation," rename the stage — do not create a custom object because the label felt wrong.
Run a sandbox load twice
First load: find duplicates, broken emails, orphan contacts. Fix the file.
Second load: validate with real users — a sales manager and one rep who will be blunt. If they say "I do not trust this number," listen.
Cutover day
Freeze edits in the old system. Export final. Import. Spot-check 20 random accounts and every open deal over $X. Keep the old export for 90 days — someone will ask.
Rushing migration to hit a board date is how you get two years of tech debt on day one. If you are mid-panic, tell us what you are importing — we have done cutovers on tighter timelines, but we will be honest if the data needs another week.