Time-Decay Attribution
Time-Decay Attribution is an attribution model that gives more credit to touchpoints closer in time to the conversion and less to earlier ones, using a decay curve.
Also known as: decay attribution, recency-weighted attribution, exponential-decay attribution
Time-Decay Attribution is a multi-touch model that weights interactions based on recency. Touchpoints that happen just before a deal closes receive the largest share of credit, while early interactions receive progressively smaller shares as they fade further from the conversion event. It encodes the assumption that recent touches matter more, which is sometimes true and sometimes misleading.
What Time-Decay Attribution Means
Time-Decay Attribution applies a decay curve, often a half-life function, so credit shrinks as touchpoints move further from the conversion date. With a seven-day half-life, a touch seven days before conversion gets half the credit of a touch at conversion, and a touch fourteen days before gets a quarter. Marketers use Time-Decay because in many sales cycles the interactions nearest the decision genuinely carry more weight, and the model rewards channels that help push deals over the line. It suits short cycles and late-stage acceleration questions better than long cycles where decisions form months before close.
How Time-Decay Attribution Works
An exponential decay function (typically half-life based) assigns weights to each touchpoint based on time-to-conversion, then normalizes the weights so they sum to 100 percent. The half-life parameter controls how quickly credit decays: a seven-day half-life is a common default but rarely appropriate for long B2B cycles. For a six-month B2B cycle, a 30 to 60-day half-life is more appropriate than the seven-day default. Tuning the half-life to your actual sales cycle is the critical implementation decision; accepting a vendor default usually produces results that look similar to last-touch.
Common Pitfalls and Misconceptions
The nuance is that recency is not the same as importance. An early webinar might have been the moment a buyer truly committed to evaluating you, yet Time-Decay will underweight it because the conversion is months later. The model embeds a strong assumption about how influence works that does not match every business. The second pitfall is leaving the default half-life when it does not match the cycle: the wrong half-life applied to a long cycle effectively erases the first half of the buyer journey from the attribution view, which is why teams using Time-Decay on enterprise B2B often end up with last-touch results in disguise without realizing it.
Time-Decay Attribution in Practice
The practitioner application is tuning the half-life to the actual sales cycle rather than accepting the platform default. Calibrate the decay to the cycle, or use a different model. For long enterprise cycles, U-shaped or W-shaped position-based models often give a more balanced view than Time-Decay tuned long, because they reward distinct milestones rather than treating influence as a smooth gradient. If you change decay parameters mid-stream, restate historical periods under the new methodology so leadership reads consistent numbers across time; switching from 7-day to 30-day half-life can make a channel look meaningfully better or worse without anything actually changing in the business.
Frequently asked questions
-
What decay rate should I use?
There is no universal answer. A common default uses a seven-day half-life, but the right rate depends on your sales cycle length. Longer cycles usually call for a slower decay so early touchpoints are not stripped of all credit. For a six-month B2B cycle, a 30 to 60-day half-life is more appropriate than the seven-day default.
-
Why choose time-decay over linear attribution?
Time-decay reflects the intuition that recent interactions tend to influence decisions more directly. If your goal is to understand what accelerates deals near the finish line, time-decay surfaces that better than equal weighting. For short cycles or late-stage analysis, time-decay is more informative than linear.
-
Does time-decay favor bottom-of-funnel channels?
Yes, it tends to credit late-stage channels like sales emails, demos, and retargeting more heavily. That can be a feature or a bug depending on whether you want a balanced view of the whole journey. Time-decay applied to long cycles often produces results that look similar to last-touch, which can undermine the whole point of using a multi-touch model.
-
Is time-decay suitable for long B2B sales cycles?
It can be, but you must tune the decay carefully. With a steep curve, touchpoints from months one and two of a year-long cycle receive almost no credit, which may undervalue important awareness-building activity. For long cycles, W-shaped or U-shaped attribution often gives a more balanced view than time-decay tuned long.
-
How is time-decay different from last-touch attribution?
Last-touch gives all credit to the final interaction. Time-decay still rewards the final touch most, but it distributes some credit backward across earlier interactions, producing a more nuanced picture. Time-decay can collapse toward last-touch when the half-life is very short relative to the journey length.
-
How do you decide between time-decay and U-shaped or W-shaped?
Time-decay treats the journey as smooth, with influence fading gradually. U-shaped and W-shaped treat the journey as having distinct milestones that deserve outsized credit. If your buyers reach clear, defined stages (first touch, lead-conversion, opportunity creation), position-based models match the reality better. If buyers move through a continuum without distinct steps, time-decay can be more natural.
-
Can you change time-decay parameters mid-stream?
You can, but it makes historical comparisons unreliable. Channel rankings shift when the decay rate changes, so a switch from 7-day to 30-day half-life can make a channel look meaningfully better or worse without anything actually changing in the business. If you change parameters, restate historical periods under the new methodology so leadership reads consistent numbers across time.