Revenue technology's boldest number belongs to Clari, whose forecast product markets "98% forecast accuracy by week two". Take the number seriously, in the way serious numbers are taken, and read it with the method held up next to the headline. We have no basis to call the claim false. The problem is that, as published, nobody can evaluate it at all, and the ways it resists evaluation are a tour of everything wrong with how this category talks about accuracy.
Four pieces are missing.
The four missing pieces
- The error metric. "98% accurate" has no standard meaning. Within a percentage band? Which band? Weighted how? Every choice moves the number by points, and the sentence names none of them.
- The horizon. A quarter's total, forecast in week two of a thirteen-week quarter, is a prediction about a number that is substantially already booked. Week-two accuracy is cheap in a way week-minus-six accuracy is not, and the claim does not distinguish them.
- The baseline. By week two, the naive forecast is strong. Simple arithmetic on the deals already in play gets most of the way unaided. The achievement worth paying for is the margin over that, and no baseline is mentioned.
- The selection. Measured on which quarters, for which customers? An average that does not say how its cases were chosen invites cherry-picking, and nothing in the claim rules it out.
None of these objections is exotic. They are the contents of a methods section, the part of any serious empirical claim that lets a reader check it. This category publishes the abstract without the paper.
Why nobody will do this to themselves
Be fair about the incentive. Revenue technology sells to revenue leaders in their own language, and that language rewards confident round numbers. A vendor who published error metrics by horizon, against a naive baseline, would be volunteering complexity into a market that punishes it, and would be calibrating a number everyone else leaves vague. The first mover pays a real cost. That is exactly why buyers cannot leave this discipline to vendors.
Four questions for procurement
Put these to any revenue forecasting vendor, ours included. Which error metric, exactly? Accuracy from which week, against the quarter's final number? What does the naive baseline score on the same data, same weeks? And which quarters went into the average, chosen by what rule? A vendor with good answers will enjoy the conversation. A vendor without them will change the subject to workflow.
Prophesee's revenue forecasts come with the methods section attached. They carry stated error metrics, accuracy by horizon and the naive baseline alongside, computed on your own pipeline before you commit to anything. Publishing the method keeps us honest, and it is hard to copy for the same reason it is rare; a methods section can be checked. Start here.