Why should buying signals decay, and how fast?

Why should a buying signal lose weight over time?

A buying signal should lose weight because it records a moment, and a buying window is a moment too. The Demand Compass treats an account's position as a reading of right now, and right now expires. The engineer who read your docs this week is telling you something urgent; the same visit seven months ago tells you almost nothing. A funding round announced last month is a live buying trigger; the same round eighteen months later is history.

Let every signal count forever and you do not have a compass. You have what we call a graveyard that glows: a market map lit up by evidence that stopped being true a long time ago. This is the Decay Rule, and the sentence we use to carry it is short enough to repeat in a meeting: old evidence is not current intent.

Let every signal count forever and you do not have a compass, you have a graveyard that glows.

The cost is arithmetic. Signals only ever get added, so a system with no decay has a score that can rise but never fall, and the order you work accounts in reflects history, not the present.

Which signals go stale fastest?

Readiness signals go stale fastest, because the readiness-to-buy axis measures internal momentum, and momentum is the part of an account that changes weekly. A docs visit this week is live; the same visit seven months ago is archaeology. Teams that let old readiness signals count at full weight chase windows that closed long ago, which is why we say decay on this axis is the difference between acting on the present and haunting the past.

The brand awareness axis decays too, but more slowly, because recognition lingers longer than urgency does. An account that attended your workshop, opened six newsletters and whose founder follows your CEO knows who you are, and it keeps knowing for a while after it stops showing up. Recognition still fades when you stop showing up, so recent engagement weighs more than old.

In both rows the signal did not become false. It became old, and a system that cannot tell the two apart presents history as current.

How fast should a signal decay?

The Decay Rule fixes the shape and leaves the speed to you: recent signals carry close to full weight, older ones fade on a curve until they are gone. Our anchors are a signal from March, which is not evidence of intent in June, and a contact who went quiet six months ago, who should not receive a sequence. We do not give a single half-life in days, and we would distrust anyone who handed you one.

The reason is that the speed depends on the signal and on your market. A pricing-page visit, a job post and a funding round are not worth the same, and they do not age at the same rate. The structure of the engine transfers to any business. The exact weights, and how quickly each one fades, are a hypothesis you calibrate against your own outcomes, the fourth step of the install and the part we earn our living tuning. Copying someone else's decay curve gives you someone else's mistakes.

How does the Demand Compass apply decay?

The Demand Compass applies decay inside the scoring step. Each account gets two scores, one per axis, each on a zero-to-ten scale and each a sum of sub-scores capped at the top of the range. The engine adds the sub-scores, applies the decay, caps the total and returns the axis. The arithmetic is deterministic: the same signals, run through the same weights with the same decay, always produce the same score.

That determinism keeps decay trustworthy. A rep can open an account and see the typed tags behind its numbers, "pricing page, three visits" or "senior security role posted." When a score drops, the tags show why, so nobody has to take the fall on faith.

We built a small version of this into our own outbound before we wrote it into the engine, and it changed our numbers. Our tool filters out inactive accounts and scores the rest by how recently they have actually done something, so we never burn a sequence on a contact who went quiet six months ago. That is our own observation, not a benchmark.

What does decay change about how you work an account?

Decay turns activation into continuous re-assignment instead of a one-time stamp. Because evidence ages, an account moves between quadrants without anyone touching it. A Cold account that catches a trigger becomes In-Market overnight, and you switch motions at once, off the slow drip and onto speed and precision. An In-Market account you warmed but did not close drifts toward Sales-Ready if the awareness took, or back to Nurture if the trigger passed.

A team that stamps an account once and runs the same motion for a year is fighting the map. Decay keeps the Market Map honest enough to act from, which is why a creeping false-positive rate is the earliest sign that signals have gone stale. Source: automate rev.ops. deployments, The Demand Compass, chapter 18

The weights age too: a signal that predicted readiness last year fades. So the weights learn offline, a proposed set runs in shadow beside the live model, and a human approves it as a new version with the old one still sitting there.

How do you check that your decay is set right?

You check it with a holdout. We hold out 5 to 10 percent of qualifying accounts at random and do not act on them. If the accounts we worked convert meaningfully better than the ones we deliberately ignored, the score is predicting something real. If they convert the same, the score, and the decay inside it, are fooling us, and the holdout says so in numbers we cannot argue with. Source: automate rev.ops. deployments, The Demand Compass, chapters 18 and 20

We tune thresholds and decay weights to hold the false-positive rate at a level the sales team can live with. The honest target is not zero: zero false positives means your threshold is so high you are missing most of the real windows too.