What is time-decay attribution?
Time-decay attribution gives more credit to the touchpoints closest to a conversion. Here is how the half-life calculation works, a worked AED example, when the model misleads, and which tools still offer it.
Time-decay attribution is a multi-touch attribution model that gives more credit to touchpoints that happened closer to the conversion and less to earlier ones. Credit usually halves for every fixed number of days back, a period called the half-life. It suits short buying cycles where the most recent touches tend to drive the decision.
How is time-decay attribution calculated?
Each touch gets a weight based on how many days before the conversion it happened. With a 7-day half-life, a touch on the day of purchase weighs 1, a touch 7 days earlier weighs 0.5 and a touch 14 days earlier weighs 0.25. Divide each weight by the total to get its share of the credit.
Time-decay attribution example
An illustrative Dubai customer buys an AED 1,500 order after four touches. With a 7-day half-life the weights add up to 1.875:
| Touchpoint | Days before purchase | Weight | Credit |
|---|---|---|---|
| Snapchat ad | 21 | 0.125 | AED 100 |
| Non-brand search ad | 14 | 0.25 | AED 200 |
| 7 | 0.5 | AED 400 | |
| Brand search ad | 0 | 1 | AED 800 |
The Snapchat ad that started the journey earns AED 100, a fifteenth of the order. Under first click it would earn all AED 1,500.
When does time decay mislead?
- It undervalues awareness: prospecting channels that start journeys always sit furthest from the sale.
- The half-life is a guess: a 7-day half-life suits quick purchases, not a three-month B2B deal. Match it to your typical time to convert.
- Google no longer offers it: Google removed time decay from Google Ads and GA4 in 2023, moving remaining conversion actions to data-driven attribution.
Related terms
Compare it with linear attribution and position-based attribution, and check your attribution window, which decides how far back the decay can reach.
How Grower helps
Grower's attribution engine de-duplicates conversions from every connected channel and app into one count, then credits each touchpoint with the model you pick: last click, first click, linear, time decay, position-based (40/20/40) or custom weights.
Frequently asked questions
What half-life should I use for time-decay attribution?
Start from your own data. Look at the typical number of days between first visit and purchase, then choose a half-life that gives early touches some credit without letting them dominate. Short ecommerce cycles often suit a week or less. Longer B2B cycles need a longer half-life or a different model, such as position-based or custom weights.
Is time-decay attribution better than last click?
It is usually fairer, because earlier touches still get some credit instead of none. It still leans heavily toward the final touches, so brand search and retargeting continue to look strong. If your main question is which channels start journeys, first-click or position-based models will answer it more directly than time decay will.
When should you avoid time-decay attribution?
Avoid it when awareness work drives your growth, such as a launch, a new market or a long B2B sales cycle. In those cases the touches that matter most happen weeks before the sale, and time decay gives them very little. Use position-based or a custom model that protects first-touch credit instead.
Run time decay next to every other model on one de-duplicated conversion count: book a Grower demo, or read marketing attribution models explained.