Why scores inflate without it
Point-based scoring is additive by default. Every email open, page view, and content download adds to the total and nothing ever subtracts. Run that model for two years and the top of your score distribution fills with long-tenured contacts whose activity stopped long ago. The threshold that once identified genuine interest now catches anyone who has been around long enough.
Reps notice before the dashboard does. They work a few of these leads, find nobody home, and start ignoring the score. Once the score loses credibility with the sales team, routing rules built on it lose their effect too, because the priority queue no longer changes what anyone works first.
Choosing a decay model
Two approaches cover most cases.
- Step decay. Engagement points drop by a fixed percentage at set intervals, for example half the value after 30 days and nothing after 90. It is easy to explain and easy to audit, which matters when marketing and sales are negotiating the threshold. - Rolling window. Only behavior inside a trailing window counts at all, and everything older falls out. This produces the cleanest read on current intent and the sharpest swings in score, so it pairs well with high-frequency inbound.
Set the window from your own data. Look at how long before a closed won deal the winning account showed its first meaningful engagement, then decay anything meaningfully older than that. Behavior that predates the typical buying window is history, not intent.
Keeping fit and engagement separate
Decay only works when the score has two components. Fit describes the account and does not age. Engagement describes the person's behavior and ages quickly. A combined score of 85 tells you nothing about whether you are looking at a perfect-fit account that went quiet or a poor-fit account clicking every email this week, and those two leads deserve completely different treatment.
Keeping them separate also makes the recycling decision obvious. When a lead's fit score stays high and its engagement score decays below the threshold, the lead belongs back in nurture rather than in a rep's queue. That single rule keeps the queue full of leads worth a call, which is what protects the conversion rates your sales forecasting inputs depend on and keeps top-of-funnel volume tied to real win rate outcomes.
Frequently Asked Questions
How fast should lead scores decay?
Tie the decay period to your sales cycle rather than to a round number. If most opportunities close within 90 days of first meaningful engagement, then engagement older than 90 days says little about current intent and should carry little or no weight. Teams with long enterprise cycles decay slower, and product-led teams with short cycles decay faster.
What is the difference between score decay and negative scoring?
Negative scoring subtracts points for disqualifying attributes or actions, such as a student email domain or an unsubscribe. Decay subtracts points purely because time has passed since the behavior happened. One judges the signal, the other judges the age of the signal, and a working model needs both.
Should firmographic points decay too?
No. Fit attributes such as company size, industry, and technology stack do not become less true over time, so decay should apply to behavior only. Split your score into a fit component and an engagement component, then decay the engagement side. Combining them into one number makes it impossible to decay correctly.
What happens if you never decay scores?
Scores accumulate until a large share of your database sits above the qualification threshold on the strength of activity from two years ago. Reps get handed leads flagged as hot that have shown no recent interest, acceptance rate falls, and the queue stops sorting by anything useful. The scoring model becomes a tenure ranking rather than an intent ranking.
Put these metrics to work
ORM builds custom revenue forecast models that turn concepts like lead score decay into prescriptive action for your team.
Schedule a Demo