Time-Decay Attribution Model Explained (and an Honest Note)

Time-decay attribution gives more credit to the touches closest to purchase — in the worked $500 example below, the final paid-search click takes roughly $290 while the first organic visit takes about $20.

Time-decay attribution is the multi-touch model that weights each touchpoint by how recently it occurred before the conversion, so later touches receive exponentially more credit than earlier ones. It is a reasonable middle ground between last-touch (which gives the closer everything) and linear (which treats every touch equally), and it is popular in tools that offer it.

TrackRev, deliberately, does not ship it as a preset — and this guide is honest about why.

First it works the maths so you understand exactly what time-decay does, then it explains the case for shipping three auditable models instead, and how to get most of time-decay’s insight from the models TrackRev does provide — each one tied to real billing revenue so it still answers which channel pays your MRR.

Why this matters for your revenue

The intuition behind time-decay is genuinely sound: a click two days before purchase probably mattered more than one three weeks earlier, and a model that reflects that feels more realistic than either single-touch extreme or a flat linear split.

For short-to-mid sales cycles, weighting by recency often does track how influence actually accrues, and that is a real reason teams reach for it.

But the same recency weighting that makes time-decay attractive also encodes a hidden assumption — the half-life — that quietly moves money between channels.

A 7-day half-life and a 30-day half-life apportion the same journey very differently, and almost nobody sets that parameter deliberately or can explain it later.

The financial risk is subtle: budget shifts based on a decay curve chosen by a tool’s default, not by a decision you made.

Awareness channels, which sit early in the journey, are systematically down-weighted, so time-decay shares last-touch’s bias toward closers, just less severely.

Understanding the model — and its buried parameter — is what lets you judge whether its answer is one you can stand behind. For the underlying channel data, see the SaaS attribution benchmarks.

What time-decay attribution is

Time-decay attribution assigns each touchpoint a weight that decreases the further it sits from the conversion, then splits the revenue in proportion to those weights. The decrease is exponential and governed by a half-life: the interval over which a touch’s weight halves.

With a 7-day half-life, a touch seven days before purchase carries half the weight of the closing touch, one fourteen days before carries a quarter, and so on.

The result is a curve rather than a rule: no touch is zeroed out (unlike single-touch models) and no touch is equal (unlike linear).

The closer dominates, the opener gets a sliver, and the middle touches sit in between according to the decay. It is the most “intuitive” of the weighted models, which is precisely why its hidden parameter is easy to overlook.

A worked example: a $500 sale

The same four-touch, 28-day, $500 journey used across this series. Time-decay is the first model here whose split you cannot do in your head — it needs the decay maths.

The four-touch journey

Organic search on day 1, newsletter on day 9, an affiliate link on day 21, and the closing paid-search click on day 28.

The gaps between touches are what drive time-decay’s weighting: the day-28 touch is the reference point, and each earlier touch is discounted by how many days before it fell.

How time-decay credits the $500

Using a 7-day half-life, the closing paid-search touch takes the largest share — about $290 — the affiliate link seven days out takes roughly $145 (half the closer’s weight), the newsletter about $45, and the discovery organic visit only about $20.

Every channel gets something, but the last two touches together take nearly 87% of the sale. Recency dominates.

The maths, step by step

Each touch gets a raw weight of 2 raised to the power of (minus days-before-purchase ÷ half-life).

With a 7-day half-life: paid search (0 days) scores 1.0, affiliate (7 days) scores 0.5, newsletter (19 days) scores about 0.15, organic (27 days) scores about 0.07. Sum the weights (~1.72), then give each touch that fraction of $500.

That yields roughly $290, $145, $45, and $20. Change the half-life to 30 days and the curve flattens toward linear; shrink it toward 1 day and it collapses toward last-touch. The half-life is the whole model.

The same sale under all five models

Time-decay sits between last-touch and linear on the identical journey — more graduated than last-touch, more recency-weighted than linear.

Reading the comparison

Time-decay and position-based both credit all four touches but disagree sharply on the opener: position-based hands organic $200 (it was the first touch), while time-decay hands it $20 (it was the oldest).

That is the models encoding opposite beliefs — position-based says discovery is special, time-decay says recency is what counts. Neither is a fact; each is an assumption, and time-decay’s assumption is tuned by a half-life you rarely see.

The strengths and the blind spots

Time-decay is a defensible model with one auditability problem that matters more than it first appears.

What time-decay gets right

For short-to-mid sales cycles, time-decay often does match how influence accrues: the touches near the decision genuinely tend to carry more weight, and crediting them accordingly — without zeroing out the earlier ones — is more realistic than either single-touch model or a flat split.

It is a sensible instinct made into a model, and in tools that offer it, it is a reasonable default for funnels where recency and conversion are tightly linked.

Where time-decay misleads

Time-decay misleads exactly where awareness matters. Because early touches are down-weighted by construction, it systematically underprices top-of-funnel channels — the podcast or content piece that created demand weeks before the sale looks minor next to the retargeting click that closed it.

For long B2B cycles, or any funnel where discovery is the hard part, time-decay inherits a softened version of last-touch’s blindness.

And its buried half-life means two teams can run “time-decay” and get materially different answers without realising they chose different curves.

Who should use time-decay

Time-decay suits specific funnels, and it is worth being clear about which.

A good fit for short-to-mid cycles

Time-decay fits when your sales cycle is short to medium and the touches near the decision genuinely do most of the persuading — many self-serve and PLG funnels look like this.

If you want to credit the closer heavily without erasing the assists entirely, and you are prepared to set and defend a half-life, time-decay is a reasonable model in a tool that offers it.

A poor fit for awareness-led growth

Avoid time-decay when discovery is the hard, expensive part of your funnel and you need to defend awareness spend — it will down-weight exactly the channels you are trying to justify.

For those situations first-touch (to isolate discovery) and linear (to credit everyone) are more honest, which is part of why TrackRev leans on that trio. See the attribution-models comparison for how they line up.

Time-decay and TrackRev: an honest note

TrackRev does not ship time-decay as a preset, and that is a deliberate design choice rather than a missing feature.

Why TrackRev ships three auditable models instead

TrackRev ships first-touch, last-touch, and linear — three models you can audit completely. For any sale, you can point to exactly why a channel got its credit: it was first, it was last, or it was one of N equal touches.

Time-decay cannot make that promise cleanly, because its output depends on a half-life that is invisible in the number and rarely chosen on purpose.

For most SaaS teams, an attribution figure you can explain to finance and defend in a budget meeting is worth more than a marginally more sophisticated curve you cannot.

Explainability over a hidden half-life

The core objection is auditability. When a channel’s revenue changes under time-decay, the cause might be a genuine shift in behaviour or just the decay curve — and you often cannot tell which.

TrackRev’s position is that at the data volumes most SaaS businesses have, a transparent model you can reason about beats an opaque one you have to trust.

A half-life you did not set and cannot see is a poor foundation for moving budget.

Approximating the recency insight

You can get most of what time-decay tells you from the models TrackRev ships. Last-touch is the recency-maximal view — it is time-decay with the half-life driven to zero — so reading last-touch already surfaces the “what closed the deal” signal.

Compare it against linear and first-touch and you can see both the closers and the creators explicitly, rather than blending them into a single curve whose shape you cannot inspect.

In practice, last-touch plus linear brackets the recency question without the hidden parameter.

When NOT to use TrackRev

If a time-decay or position-based preset is a hard requirement for your reporting, or you need a black-box data-driven model with a large training set, TrackRev is not the tool — it stands behind three auditable models on purpose.

If your billing is not on Stripe, Paddle, Polar, or Lemon Squeezy, the revenue join does not apply, and if you sell physical products or run a CRM-led enterprise motion, a specialist tool fits better.

TrackRev is for SaaS teams that would rather have attribution they can explain than attribution they cannot.

The stack maths is the consolidation case: Bitly Growth (~$35/mo) plus Rewardful Starter (~$49/mo) is $84+/month for two tools with two definitions of a sale, while TrackRev is $39/mo for link tracking, attribution, and affiliates on one.