Let’s be honest — moving data into Salesforce isn’t the exciting part of a CRM project. Nobody celebrates a “clean import.” But if you’ve ever tried to migrate millions of records, you know it’s where things can fall apart fast.

We’ve run plenty of these at Revenue Ops, and every time, the lesson is the same: big data moves are about patience, planning, and respect for scale. You can’t wing it.

Why large-volume migrations are different

If you’re moving 50,000 records, you can get away with Data Loader and a long afternoon.
But when you’re working with millions of Accounts, Contacts, or Opportunities, things start to break down. Salesforce itself calls this “Large Data Volume” territory (Salesforce LDV Guide), and it changes the rules completely.

You’ll see things like:

• Slow everything — list views that crawl, reports that time out, sharing recalculations that take hours.
• Record locking — too many child records tied to one parent (the dreaded “skew”).
• Automations blowing up — flows, triggers, and processes firing thousands of times in one batch.
• Dirty data showing its face — duplicates, missing values, and relationships that never matched up right.

The fix? Treat it like a real project — not a file import.

Step one: get clear on what actually matters

Start with why. Why are you moving the data at all? Are you consolidating systems, rebuilding your data model, or finally cleaning up years of technical debt?
Once you know that, decide what’s worth moving. Not every record deserves a spot in the new org. Archive the noise.

Then write down what “success” looks like:

• Which objects are included?
• How clean should the data be when it lands?
• What’s your acceptable downtime or risk level?

If those questions make people squirm, that’s a good sign you need to slow down before you start.

Step two: clean before you move

You can’t polish garbage once it’s in Salesforce.
Do your cleanup ahead of time.

• De-dupe the source.
• Fill in missing IDs.
• Get rid of half-baked records nobody’s touched in years.

Build a simple data mapping sheet — source field on one side, Salesforce field on the other, and notes for any transformation rules.
It’s boring work, but it’s the foundation.

Step three: choose the right tools

For smaller batches, Data Loader works fine. For the big stuff, go with Bulk API 2.0.
If you’ve got multiple systems or weird transformations, use an ETL tool — MuleSoft, Talend, or Informatica.

For historical or never-touched data, don’t force it into standard objects. Use Big Objects or external storage (Big Objects Overview).
And if your automations are heavy, turn them off during the load. You’ll save hours and headaches.

Step four: test like you mean it

Don’t do a 500-record “test” and call it good. Load a real slice — 10 to 20% of your total.
Run reports. Open list views. Click through related lists.
If Salesforce starts to lag, fix it now. You won’t have time once it’s live.

Step five: plan your cutover

You have two options:

• Big bang — move everything at once (risky but fast).
• Phased — migrate by business unit, region, or timeframe (safer but slower).

Whichever you choose, have a rollback plan. Backups. Screenshots. Something.
No one wants to explain missing Contacts on a Monday morning.

Step six: validate and tune

Once the data’s in, check everything.

• Record counts match?
• Relationships intact?
• Sharing and ownership correct?
  Run reports side-by-side with your source systems.
  Watch how users interact with it. If they complain about performance, listen — it’s usually valid.

Quick best-practices we live by

• Don’t dump everything. Archive what you can.
• Break up large loads into smaller chunks.
• Disable non-essential automations during migration.
• Keep your mapping doc updated — you’ll need it later.
• Communicate like crazy. Surprises are what break projects.

Salesforce has a solid best-practices PDF on this stuff if you want the technical details.

What we’ve learned at Revenue Ops

After years of helping teams move massive datasets, the best advice we can give is this: respect the data.
Don’t rush it. Don’t overcomplicate it.
Build a plan that’s technical enough for engineers and clear enough for executives.

At Revenue Ops, we help teams make migrations smooth, predictable, and clean — so you end up with a Salesforce org you can actually trust.

If you’re staring down a big migration and want a sanity check, reach out. We’ll walk you through what works and what doesn’t — no fluff, no sales pitch.

Final thought

Large-volume migrations are rarely fun, but they’re the difference between a cluttered CRM and a system that truly supports your business.
If you plan it right, test it properly, and bring the right people to the table, you’ll only have to do it once.

Do it wrong, and you’ll be “fixing” your migration for the next year.