Essay · Architecture
Why we put the CRM, the data and the outreach on one database
Every revenue team we met was paying four companies to keep four copies of the same customer. This is what we learned building the alternative.
When we asked sales leaders to draw their stack,1 almost every drawing had the same shape. A CRM in the middle. A data vendor feeding it. A sequencer beside it with its own copy of every contact. A dialer with a third copy, and a chat widget on the website with a fourth.
None of these tools is bad. The trouble is between them. Each one has its own idea of who a customer is, and someone on the team spends Fridays reconciling those ideas.
What sync really costs
Sync looks free on the pricing page. It shows up later: as duplicate records, as a prospect who gets a cold email the day after they booked a demo, as a forecast built on stages nobody updated.
“The most expensive integration in a sales stack is the one a person does by hand every Friday.”
So we started from the record. One row per company and per person, written to by every part of the product: the website visitor, the enrichment, the reply, the call, the meeting.
- Apr
- May
- Jun
- Jul
- Aug
- Sep
Every enrichment one team pays for is cached for the next. Illustrative data.
The caching dividend
A shared database has a side effect we didn’t plan for. When one customer enriches a company, the next customer who needs it gets it from our own store.2 The cost of a credit falls as the database grows.