A deal flow CRM has one field that returns more than all the others combined and three that get skipped no matter how the form is designed. Knowing which is which before you roll one out saves a year of arguing about data quality, because the skipping is not laziness: it is a rational response to a field that costs something to fill and returns nothing to the person filling it. This page sets out the one worth insisting on and the three worth dropping.
The one that pays: where did this come from?
One dropdown at entry, three or four options, and it is the highest-return field in the record. Without it, no firm can say in a year which channel is worth the time it costs, which means the resourcing argument gets settled by whoever is most confident in the room. With it, the argument is arithmetic. Insist on this one and be relaxed about the rest.
Skipped one: estimated deal size at first contact
Nobody knows, everybody guesses, and the guess then anchors the pipeline value for months. It reads like rigour and produces the opposite. Better to leave it empty until there is a figure worth recording, and to value the book on an average rather than on a column of first impressions.
Skipped two: probability percentages
Asking an associate for a percentage produces a number shaped by how the last conversation felt. Stages are a better instrument because they are observable: an NDA is either back or it is not. A close rate measured across stages is more honest than a hundred individual estimates averaged together.
Skipped three: the long-form note nobody reads
A free-text field that expects a paragraph gets a sentence or nothing. A one-line note field gets a line, every time, which is more useful in aggregate. Design for the note somebody writes on a bad Thursday rather than the one they would write with an hour free.
Questions people ask about deal flow crm
What if we want the estimated size anyway?
Make it optional and never use it for the pipeline value. Value the book on an average and a measured close rate instead.
Are probability percentages ever useful?
Rarely, and mostly as a comfort. Stages are observable, which makes a close rate measured across them a better instrument.
How do we get people to fill in the source?
Make it a dropdown with four options and make it the only required field. Everything else can be optional and the record still works.