Predictive Attribution Model
What is Markov Attribution? Think of every customer journey as a path:
LinkedIn Ad → Website → Demo Request → ConversionThe Markov model looks at all these paths and asks: “If we removed this channel from the journey, how much would conversions drop?”
If conversions fall significantly when a channel is removed → that channel gets more credit.
If conversions barely change → the channel gets less credit.
This approach captures the true incremental impact of each touchpoint.
The Problem with Standard Markov
Without adjustments, Markov models often over-credit very common channels (like “Direct” or “Brand Search”) simply because they appear in nearly every journey. That means:
“Direct” might look like your best channel (because it’s everywhere).
Niche but powerful touchpoints (like a webinar invite or specific ad campaign) get under-valued.
The HockeyStack Twist: Inverse Frequency Weighting
We fix this with Inverse Frequency Weighting (IFW).
Common touchpoints that appear in almost every journey (like “Direct”) get down-weighted.
Rare but meaningful touchpoints (like “Event Sponsorship” or “Cold Email”) get up-weighted.
The result:
A more balanced, fair attribution model.
Clearer visibility into the channels that actually drive conversions, not just the ones that show up most often.
Important Limitation to Know
Due to model complexity, the predictive attribution model itself is not customizable, and can't be accessed directly through the UI to see exactly what’s happening under the hood.
Example Imagine two channels:
Direct: present in 80% of journeys.
Webinar Invite: present in 10% of journeys.
In standard Markov:
Both get similar credit if they show up before conversions.
With IFW:
Direct gets reduced credit (because it’s always there, not discriminative).
Webinar Invite gets boosted (because it’s rarer and more influential when it appears).
This lets you uncover hidden ROI drivers that traditional models miss.Why This Matters for You
See past the obvious: Don’t get fooled by overused channels.
Identify true growth levers: Spot the campaigns, events, and ads that really change outcomes.
Invest with confidence: Allocate budget to the activities that incrementally drive pipeline, not just the ones that show up the most.
Summary
Markov Attribution with Inverse Frequency Weighting gives you the most realistic view of how your marketing channels contribute to revenue:
Markov finds incremental impact.
IFW balances common vs. rare touchpoints.
Together, they provide the clearest possible map of where your marketing dollars actually work.
Last updated