Time decay attribution model
Credits touchpoints more heavily the closer they occur to the conversion event, using an exponential decay function. Reflects recency, but risks undervaluing the awareness channels that started the deal.
What is the time decay attribution model?
The time decay attribution model assigns credit to touchpoints using an exponential decay function applied backwards from the conversion event. Touchpoints that happened one day before conversion receive significantly more credit than touchpoints that happened thirty days before, which in turn receive more credit than touchpoints from three months ago. The total credit across all touchpoints still sums to 100% of the conversion value.
The underlying assumption is that recent interactions are more causally connected to the conversion decision than older ones. This is often true in short-cycle, high-intent scenarios. A prospect who clicked three different pricing pages in the 48 hours before requesting a demo was probably more influenced by those clicks than by a blog article they read four months ago.
The model was popularised by Google Analytics and remains common in e-commerce and direct-response advertising, where purchase cycles are short and recent behaviour is a reliable predictor of purchase intent. Its fit for long B2B cycles is less clear, because the seed-planting interactions at the top of the funnel can be months removed from the eventual deal creation.
How credit is distributed
The standard time decay implementation uses a half-life of seven days. A touchpoint that occurred seven days before conversion receives half the credit of a touchpoint that occurred on the conversion date. A touchpoint fourteen days out receives one quarter, and so on. The weights are then normalised to sum to 100%.
| Touchpoint | Days before conversion | Relative credit |
|---|---|---|
| Organic blog visit | 60 days | ~2% (lowest) |
| Webinar attendance | 21 days | ~11% |
| Nurture email click | 7 days | ~22% |
| Demo request (paid search) | 0 days (conversion) | ~65% (highest) |
Compare with linear attribution (25% per touchpoint in this example) and last-touch (100% to the demo request). Time decay sits between them on the recency spectrum.
When to use time decay attribution
Time decay is best matched to:
- Short sales cycles of one to four weeks where recent activity is genuinely the strongest predictor of conversion
- E-commerce and direct-response contexts where the intent signal degrades quickly and the buying window is narrow
- Retargeting analysis where you want to confirm whether a recent ad impression was the accelerant it appears to be
For B2B SaaS deals with three-month-plus sales cycles, time decay systematically undervalues the content marketing and brand touchpoints that initiated and nurtured the relationship. In those contexts, pair time decay with U-shaped attribution to ensure both early and late touchpoints receive structured credit.
Pros and cons
Pros
- Accounts for recency: touchpoints close to conversion get appropriately higher credit
- More realistic than linear for short cycles where recent activity drives decisions
- Preserves multi-touch credit distribution, unlike single-point models
- Intuitive for sales teams: the actions that closed the deal are credited most
Cons
- ✕Undervalues awareness and top-of-funnel content that starts the relationship
- ✕Decay rate is an assumption: too steep and it becomes close to last-touch
- ✕Poor fit for long B2B cycles where first contact can be months before close
- ✕Can discourage investment in high-intent mid-funnel content that takes time to convert
How AttriByte handles time decay attribution
AttriByte applies a configurable half-life to the time decay model so you can match the decay rate to your actual sales cycle length rather than accepting a hard-coded default. All calculations run against the same cookieless identity-resolved event stream as the other five models.
Because the identity layer stitches sessions across devices and time gaps, time decay in AttriByte operates on the real elapsed time between interactions, not the time between browser sessions. A prospect who went dark for three weeks and then returned via a different device will have all those interactions correctly dated and weighted.
Comparing time decay to linear side-by-side quickly shows which channels are being disproportionately credited or undercredited based on timing alone. That comparison is a direct input to recency-based campaign scheduling decisions. Read more on the product page.
Explore other models
Run time decay alongside five other attribution models
AttriByte calculates all six models in parallel on your warehouse data. Adjust the decay rate to match your sales cycle.