How the time decay attribution model works, when to use it, and where it falls short

Think about how most people find an apartment. They might talk to their friends about the best neighborhoods to live in, browse listings for a few weeks, save a handful of favorites, tour two or three in person, and then do one final walkthrough right before they sign the lease. If you asked that renter what actually sealed the deal, they'd probably point to that last walkthrough. But the listing they saved three weeks earlier is the only reason they showed up to tour it in the first place.

That's the logic behind a time decay attribution model. It's a way of assigning credit to every step in a customer's decision, while still recognizing that the moments closest to the finish line usually carry more weight. For marketers trying to figure out where their budget is actually working, the attribution model you choose to measure that journey has real consequences. Pick the wrong one, and you'll end up pouring marketing spend into whatever channel happened to be standing there at the final conversion, while the earlier marketing efforts that built the customer's interest in the first place go largely unrecognized. Get it right, and you'll have a much clearer (though, not perfect) read on which marketing channels are pulling their weight across the entire customer journey.

Key takeaways

• A time decay attribution model assigns more credit to touchpoints that happen closer to a conversion event, based on the idea that recent interactions have a more significant impact on customer behavior than earlier ones.
• The model runs on a half life formula, most often set to seven days, so a touchpoint from a week before conversion earns half the credit of one that happens on conversion day.
• It's one of several attribution models, alongside a linear attribution model and a position based attribution model, and each one assigns credit across the customer journey differently.
• A time decay attribution model tends to work best for long sales cycles, nurture campaigns, and any sales funnel where multiple touchpoints build customer interest over several weeks.
• The model can undervalue early interactions like brand awareness campaigns, since initial awareness rarely happens close enough to the final touchpoint to earn much credit.
• Like other attribution models, it depends on being able to track a customer across every touchpoint in their journey, which is a harder ask than it used to be.
• Recency isn't the same as impact, so marketers should pair a time decay attribution model with a measurement approach that can confirm which marketing efforts are actually driving sales.

What is a time decay attribution model?

A time decay attribution model is a type of multi touch attribution that assigns credit for a conversion across every touchpoint a customer interacts with, weighted by timing. The final touchpoint before conversion gets the most credit, and each earlier touchpoint gets progressively less. It sits alongside other attribution models like a linear attribution model, which splits credit evenly, and a position based attribution model, which favors the first or last interaction over everything in the middle.

The idea behind a time decay attribution model is fairly simple: the closer a touchpoint sits to the final conversion, the more likely it played a meaningful role in that customer's decision making process. The assumption is that a paid search ad someone clicked right before checkout probably had more influence on that specific sale than a social media ad they saw a month earlier. Whether that assumption holds up is worth questioning, and we'll get into that later, but it's the reasoning this attribution model runs on.

How a time decay attribution model works

A time decay attribution model runs on a half life formula, similar to the kind you might remember from a high school science class. Instead of measuring how a physical substance breaks down over time, this half life formula measures how quickly a touchpoint's influence fades as time passes between that interaction and the final conversion.

The half life formula, broken down

Most analytics tools default to a seven day half life, though this default half life isn't a hard rule. Here's what it looks like when you break down how much credit each touchpoint earns:

• A touchpoint on the day of conversion earns full credit.
• A touchpoint seven days before conversion earns roughly 50% of that credit, since it sits exactly one half life away.
• A touchpoint 14 days before conversion earns roughly 25% of that credit, or two half life periods removed.
• A touchpoint 21 days before conversion earns roughly 12.5% of that credit, three half life periods out.

Each earlier touchpoint keeps losing half of its remaining value for every additional half life period that passes, which is why late stage interactions consistently earn more weight than early ones. So if Jess sees a display ad 21 days before she buys, clicks a retargeting email 10 days later, and finally converts after clicking a paid search ad on day zero, the paid search ad would carry the most weight, the email would carry a moderate amount, and the display ad would carry the least credit of the three. All three touchpoints still get some credit, and none of them get shut out the way they would under last click attribution, which assigns all the credit to that final paid search ad and nothing to the earlier touchpoints that came before it.

Adjusting the default half life

A seven day half life makes sense for a sales funnel that closes fast, but it doesn't fit every business or every set of marketing channels. A B2B company with a six month sales cycle would see almost all of its early interactions decay down to nearly nothing under a seven day setting, even though those early touchpoints might be exactly what built the customer's initial interest. Most analytics tools and other attribution models that support time decay will let you extend the half life to match your actual sales cycle, whether that's 14 days, 30 days, or longer. The right setting depends on how much time it typically takes your customers to move from initial awareness to a final conversion point.

Time decay attribution model vs. other attribution models

A time decay attribution model is just one of several ways to assign credit across a customer's decision making process. Here's how it stacks up against the other common attribution models.

None of these marketing attribution models are wrong, exactly. They're just built around different assumptions about how much weight to give recent interactions versus earlier ones. The right attribution model for your business depends on how long your sales cycle runs and how many marketing touchpoints typically factor into a customer's decision making process before they convert.

When a time decay attribution model makes sense

A time decay attribution model tends to earn its keep in a few specific situations. It's worth considering if your marketing strategies and sales funnel look like any of the following:

• Long sales cycles. B2B software, financing, and other high-consideration purchases often involve weeks or months of research across multiple touchpoints. A time decay attribution model supports long sales cycles by recognizing every step along the way, rather than crediting only the final one.
• Nurture campaigns. If your marketing efforts rely on email sequences, retargeting, and repeated touchpoints to slowly build customer interest, a time decay attribution model reflects that gradual buildup better than a model that only rewards the first or last interaction.
• Multiple marketing channels working together. Businesses running paid search, social media ads, and email marketing campaigns in tandem benefit from an attribution model that tracks those marketing efforts across various marketing channels instead of assigning credit to just one.

Where a time decay attribution model falls short

No attribution model gets everything right, and this one has three real limitations worth understanding before you build your marketing strategy around it.

It undervalues the top of the funnel

The same half life formula that makes a time decay attribution model useful also makes it biased against early interactions. A brand awareness campaign or an initial interaction from a month before conversion will almost always score low in this model, not because it didn't matter, but because the model assumes touchpoints lose relevance the further they sit from the final conversion. For businesses that rely heavily on brand awareness campaigns to fill the top of the sales funnel, this can create a skewed picture of which marketing channels deserve more credit, and it can lead to under-investing in exactly the channels that generate future demand.

It depends on tracking that's getting harder to trust

A time decay model, like every other multi touch attribution model, needs a fairly complete record of the customer interactions that make up a customer's entire journey. That record depends on cookies, pixels, and other tracking methods that browsers and devices are increasingly designed to block. As tracking gets harder, the customer journey these models can actually see gets shorter and less complete, which means the credit they're assigning is only ever as good as the data behind it. This isn't a flaw unique to time decay attribution, and it shows up in other attribution models too, but it's worth factoring into how much confidence you place in the results.

Recency isn't the same as impact

A touchpoint that happens right before a conversion event isn't necessarily the touchpoint that caused it. A customer who already planned to buy might click a retargeting ad only because it was higher on the search results, and that ad would still earn significant credit under a time decay attribution model, even if it changed nothing about their decision. A time decay model can tell you which touchpoints were present near the end of a customer's journey. It can't confirm which ones actually moved the needle, and that gap matters more the bigger your marketing budgets get.

Where Prescient comes in

Attribution models like a time decay model give you a view of the touchpoints in a customer's journey, but they can't confirm which of those customer interactions actually drove the sale versus which ones simply showed up along the way. That's a different question, and it's the one marketing mix modeling is built to answer. Prescient's platform looks at marketing performance holistically, incorporating media and non-media factors into a more data driven view of which marketing efforts are truly driving results, rather than relying on assumptions about how much weight a touchpoint's timing should carry.

If you're trying to figure out whether your current attribution model is giving you the full picture, we'd be glad to walk you through what a more complete view of your marketing channels could look like. Book a demo to see how it works.

FAQs

How is a time decay attribution model different from a linear attribution model?

A linear attribution model splits credit evenly across every touchpoint in the customer journey, regardless of when each one happened. A time decay attribution model instead gives more weight to touchpoints closer to the final conversion and less weight to earlier ones. If a customer's journey includes an early social media ad and a late paid search ad, a linear attribution model would treat them as equally influential, while a time decay attribution model would give the paid search ad more credit.

Can marketers customize the half life in a time decay attribution model?

Yes. Most analytics tools default to a seven day half life, but that setting can usually be extended to match a longer sales cycle. A business with a shorter path to purchase might keep the default half life, while a B2B company with a months-long sales cycle would typically stretch it out so early touchpoints aren't unfairly discounted.

Is a time decay attribution model a good fit for B2B marketing?

It can be, especially for B2B companies with long sales cycles and multiple touchpoints across email, paid search, and other marketing channels. The key is adjusting the default half life so it reflects how much time deals actually take to close, since a setting built for a fast consumer purchase won't fit a six month enterprise sales cycle.

What's the difference between a time decay attribution model and last click attribution?

Last click attribution assigns all the credit for a conversion to the final touchpoint and ignores everything that came before it. A time decay attribution model still gives the final touchpoint the most credit, but it also recognizes earlier touchpoints, just at a reduced weight. For businesses with multiple touchpoints across the customer journey, that difference can meaningfully change which marketing channels look effective.