Which CRM fields should be required, and how do you stop duplicate and dirty data?

Require a field only if you can name the report it appears in, and fix duplicates at the source that creates them rather than merging the same records every quarter. Both halves of that matter. Over-requiring produces fast garbage. Merging without fixing the source is a chore that never ends.

Deciding what to require

Start from the reports leadership asks for and list the fields each one depends on. That list is your required set. Everything else is optional until somebody can name the report it feeds.

Then check who fills each one in. If a required field only benefits another team, a rep will type anything to get past it. Capture it automatically from the form, the integration or the email signature instead, or drop the requirement. Required fields that people resent are how you get contacts named "asdf" with a phone number of 5555555555.

Picklists over free text

Any field you intend to group a report by should be a dropdown. Industry, lead source, closed lost reason, state, country. Free text on a reporting field guarantees forty variations of the same answer. Normalize state and country values during migration and constrain them afterwards, because a list containing Texas, TX and Tx will never group correctly. Here is a clean state and abbreviation reference if you need one for the mapping.

Keep picklists short and owned by one person. A dropdown with sixty options that anyone can add to is free text with extra steps.

Duplicates, handled by source

Merging duplicates feels productive and solves nothing if the thing creating them is still running. Find the source first. In practice it is almost always one of four:

• Imports. A spreadsheet loaded without a match key, or matched on name instead of email. Always import against a unique key.
• Forms. Someone submits with a personal address one time and a work address the next. Progressive profiling and consistent field mapping reduce this.
• Integrations writing on the wrong key. The most common culprit and the hardest to spot. A connected system creates a new record because it matched on an internal ID the CRM does not hold. Fix the match logic, not the output.
• Manual creation. A rep who cannot find an existing record makes a new one. That is usually a search or permissions problem.

Once the source is fixed, run the merge, then schedule a recurring check so you find out in weeks rather than years. Email validation belongs in the same pass, since dead addresses damage deliverability as much as duplicates damage reporting. Our email list cleaning process covers that. Deeper reporting and data work sits in our integration services.

Related: What is a CRM data migration? | What data and analytics services do you offer? | My data lives in 5 different tools. Can you help unify it?