A rep needs to close out a Friday call, Salesforce demands six fields, and the rep types "TBD" into all of them. The rule fired. The data is still junk. That's the usual outcome when validation is built to block people instead of protect a number.
Why Validation Rules Go Wrong
Most orgs end up in one of two places:
- Too loose. Nothing is required until Closed Won, so the forecast runs on guesses for 80% of the deal's life.
- Too strict. Every field is required at creation, reps feel policed, and they fill in placeholder values to get past the screen.
Both fail the same way. The pipeline report looks complete, and nobody trusts it at month-end close.
Start With the Number, Not the Field
Pick the metrics leadership actually acts on: weighted pipeline, forecast category, close date, deal size. Then work backward to the two or three fields each one depends on. Those fields get a rule. The other forty don't.
A Validation Approach That Holds
- Tie rules to stage, not to creation. Require a close date and amount when a deal reaches qualification. Require a next step and decision maker when it reaches proposal. Reps provide data when they have it.
- Block bad dates outright. A close date in the past is never right. It's the cheapest rule to write and the one that cleans up the forecast fastest.
- Write error messages that say what to do. "Enter a close date after today" beats "Required field missing." Name the field and the fix.
- Reject placeholder values. Check for "TBD," "N/A" and zero amounts on required fields where they defeat the purpose.
- Add a bypass for integrations and migrations. Use a custom permission so data loads and ERP syncs don't fail on rules meant for humans, and log who has it.
- Track rule fires. If one rule triggers hundreds of times a week, the process or the page layout is wrong, not the reps.
- Review twice a year. Retire rules nobody can explain, and test the rest after any stage or picklist change.
What Good Looks Like
Pipeline data hygiene improves without a cleanup project. Reps hit a rule only when the data genuinely matters, and the forecast call stops opening with "is this number real?" You can drill down from the rollup to a deal and trust what's on the record.
Our Take
Validation is a forecasting control, not a compliance exercise. Fewer rules, tied to stages, with clear messages, outperform a long list that reps learn to game. If you'd like a second set of eyes on your org, we'll review your active rules, flag the ones driving workarounds, and show you which fields your forecast depends on.