Fact-based news without bias awaits. Make 1440 your choice today.

Overwhelmed by biased news? Cut through the clutter and get straight facts with your daily 1440 digest. From politics to sports, join millions who start their day informed.

Sign up now!

Welcome back!

We hope everyone’s ready to start the week with some fresh marketing strategy! Today, we’re leveling up and looking at a more *sophisticated* model, called time decay attribution. Don’t worry though— we’ll keep it light on the math and explain everything simply along the way.

Let’s dive in! 🏃‍♂️🚀

News

Tools + Productivity

Strategy

What is it?

We’ve talked about giving different credit weights to different stages of the customer journey, but the models we’ve covered so far simply assign credit to touchpoints at the beginning, end, or equally throughout.

With time decay, more credit is given to marketing interactions that happen closer to when a customer makes a purchase or converts, based on the idea that recent interactions are more influential.

So what’s the difference between time decay and last touch? Rather than just lumping all the credit into one touchpoint, credit is still distributed across multiple touchpoints, but with more weight to the latter ones.

First, we should go over some of the principles that underpin this attribution model:

temporal proximity: touchpoints closer to conversion are more influential

multi-touch attribution: all relevant interactions in the customer journey are considered.

weighted distribution: all touchpoints get credit, but with decreasing weight as you move back in time.

decay function: the exponential mathematical formula used to calculate the decreasing value of touchpoints over time.

This typical decay function can be applied to time decay attribution models, where W(t) is the weight at time/touchpoint t, lambda is the decay constant, and t is the time difference between the touchpoint and conversion.

What’s to love?

While simple models are helpful when resources are limited and the brand is smaller, a pretty good rule of thumb for marketing and life is that reality is complicated.

That’s why time-decay models, although more convoluted, are closer to the true value of the various touchpoints as customers progress through the sales funnel, meaning the subsequent resource allocation decisions are better.

All touchpoints are acknowledged, but recent interactions are emphasized, creating a middle ground for the basic attribution models we discussed in past editions.

The glaring issue with time decay attribution is that it requires sophisticated tracking and data analysis capabilities, and heavily relies on accurate and comprehensive data collection across all channels.

This means inconsistencies or gaps in data can significantly impact the model’s effectiveness.

Who should use it?

This model is particularly great for B2B companies with moderate to long sales cycles, as well as those using various marketing channels and wanting to understand their relative impact over time.

Businesses with complex, nonlinear customer journeys may find that time decay attribution may not accurately reflect the cumulative impact of all touchpoints.

More importantly, organizations that can’t afford the robust systems required for tracking and analyzing customer interactions across multiple channels certainly would find it difficult to implement this model.

Implementation in action

First comes tracking interactions. Each interaction, from ad clicks to email opens to website visits to social media engagements, is logged with a precise time stamp.

Then, all recorded interactions are sorted chronologically, creating a timeline of the customer’s journey from their first engagement to the final conversion.

The decay function is applied to each interaction to calculate a weight based on its proximity to the conversion.

The total conversion value (100%) is distributed among all touchpoints, so each touchpoint receives a portion of the credit based on its calculated weight.

If we have, say, 3 touchpoints before a conversion:

Touchpoint A: 14 days before

Touchpoint B: 7 days before

Touchpoint C: 1 day before

And we use a half-life of 7 days (λ ≈ 0.099, from λ = ln(2) / 7) 

*Side note: the half-life represents the amount of time it takes for the weight/value of a touchpoint to decrease by half)

We find the weight of each touchpoint by plugging into the decay equation:

• Weight A = e^(-0.099 * 14) ≈ 0.25

• Weight B = e^(-0.099 * 7) ≈ 0.50

• Weight C = e^(-0.099 * 1) ≈ 0.91

And then normalize these weights to get the final attribution percentages:

Total = 0.25 + 0.50 + 0.91 = 1.66

• Touchpoint A: (0.25 / 1.66) * 100% ≈ 15%

• Touchpoint B: (0.50 / 1.66) * 100% ≈ 30%

• Touchpoint C: (0.91 / 1.66) * 100% ≈ 55%

Now that we have the weights of each touchpoint, we can apply the same principles from first touch and last touch to allocate resources towards campaigns accordingly.

Looking Ahead

We’re almost to the end of the month! Now that we’ve covered arguably the hardest attribution model to understand and even more so, implement, we can finish off strong with position-based attribution.

Stay tuned!

Attribution models

Week 1: Last-touch attribution
Week 2: First-touch attribution
Week 3: Linear attribution
Week 4: Time-decay attribution (this week!)
Week 5: Position-based attribution

Stay tuned!

REFERRAL PROGRAM

Hey! Don’t keep us a secret! Refer your friends and start earning rewards! Share this link with your friends: https://beyond9to5.beehiiv.com/subscribe?ref=PLACEHOLDER

If you get just 3 friends to subscribe you’ll get our exclusive sticker pack OR refer more and earn a trip to Vegas!

Your Progress