CRM as a Record

A CRM is bought to remember what happened with customers and then used to evaluate the people entering the data. Those two purposes are incompatible, and the second destroys the first.

What it is for

Remembering. Who said what, when, what was agreed, what is outstanding. Nobody holds this across a hundred relationships and two years, and a system does it perfectly.

Continuity. Somebody leaves and the relationship does not.

And pattern. Which kinds of deal close, where they stall, how long each stage takes.

All three depend on the record being complete and honest.

What it becomes

A scoreboard. Calls logged, activities per week, pipeline value updated by Friday.

The moment a number in the system determines somebody's standing, the number stops describing reality. Not through dishonesty — through ordinary self-protection under measurement, the same mechanism as taking fewer breaks to avoid appearing idle.

Pipeline value inflates. Stages get advanced early. Activities that happened get logged and activities that did not sometimes do too.

And the pattern data becomes unusable, which was the expensive thing the system was bought for.

The specific failures

Activity targets. Twenty calls a week produces twenty calls, of which some exist to be counted.

Stage-based forecasting on self-reported stages. The forecast is a report of what salespeople think you want to hear.

And close-date hygiene rules, which move dates rather than change outcomes.

Each measures something real and each is gamed within a quarter, because the person entering the data is graded on it.

How to keep the record honest

Separate what is measured from what is entered.

Judge on outcomes — closed business, retention, whatever the actual result is. Those come from elsewhere in the system and cannot be typed in.

Never on entry volume. The moment activity counts, the log becomes a performance.

Make the record useful to the person keeping it. Adoption is predicted by whether the entrant gets something back, and a salesperson who can see their own history is entering data for themselves.

And ask for less. A short record kept honestly beats a long one filled to a standard.

The question before buying one

"What decision will this change?"

If the answer is knowing what was said to whom, a CRM is exactly right and the narrow version will do.

If the answer is visibility into whether the team is working, that is a management question being answered with a purchase, and the CRM will supply a number that means nothing within two quarters.

The short version

• A CRM exists to remember, provide continuity and reveal pattern, and all three need the record complete and honest
• The moment a number in it determines somebody's standing, the number stops describing reality
• Activity targets, stage-based forecasting on self-reported stages, and close-date rules are each gamed within a quarter
• Judge on outcomes that come from elsewhere in the system and cannot be typed in, never on entry volume
• Make the record useful to the person keeping it, and ask for less — a short honest record beats a long compliant one
• If the reason for buying is visibility into whether people are working, the number will mean nothing within two quarters

For additional context on this topic, see Polygon.