Field notes on revenue operations.

Most forecasts are produced by negotiation up a management chain, then defended as arithmetic. A defensible forecast is one where every delta from last week is attributed to a specific deal and a specific event. Here is how to build one.

Most RevOps functions climb the same five stages in the same order, and most stall at the same one. Knowing which stage you are in tells you what to build next — and, more usefully, what not to.

Hygiene campaigns fail because they are framed as tidying. Reframed as the input layer of the forecast — with the cost of each defect made visible — the same work gets done and stays done.

Before a revenue team acts on what an AI system says, the claim should meet a standard. Here is a practical one: provenance, mechanism, confidence, counter-evidence, and a threshold below which the system says nothing at all.

Autonomy is not a switch, it is a dial set per action type. Risk-tier by reversibility and audience, start where a mistake is cheap, and earn the next level with evidence rather than a meeting.

Replacing the CRM almost never fixes what people blame the CRM for. The missing piece is a layer above it that knows what an account is across every system, and what happened to it in order.

Response time gets reported as a marketing SLA and optimised as an average, which hides exactly the cases that cost money. Measure the tail, fix the routing, and treat the first reply as part of the sales process rather than an acknowledgement of it.

By the time a renewal is on the calendar, the decision is usually made. The signals that predicted it were in support, usage, org changes and sentiment — scattered across systems nobody joined together.

Most pilots die of ambiguity rather than failure — nobody agreed in advance what success looked like, so the result was whatever the loudest person said it was. Here is the structure that avoids that.