Table of contents
Most attribution models start by deciding a rule for splitting credit, then apply it. Markov chain attribution does the opposite. It starts with your actual conversion paths and asks a simple, almost surgical question of each channel: how many conversions would we lose if this one vanished? That question is called the removal effect, and it is the heart of the model.
What is Markov chain attribution?
Markov chain attribution is a data-driven model that maps every customer journey as a sequence of states (channels) and calculates the probability of moving from one to the next. It assigns credit to each channel using the “removal effect”: how much the overall probability of conversion drops when that channel is taken out of the system.
A Markov chain is, formally, “a model describing a sequence of possible events in which the probability of each event depends only on the current state”. In attribution terms, the “current state” is the channel a buyer is on, and the model only cares about where they go next, not the entire history behind them.
How it works
There are three steps:
- Build the graph. Turn all your conversion paths into a map of states, from a “start” node through each channel to either “conversion” or “null” (no conversion). The transition probability between two channels is the number of paths using that step divided by all paths leaving that node.
- Calculate the removal effect. Remove one channel from the graph and recalculate the probability of reaching a conversion. The percentage drop is that channel’s removal effect. Repeat for every channel.
- Normalise. Because removal effects do not naturally add up to 100%, you scale them proportionally so they sum to one, then multiply by your actual conversions to get each channel’s credit.
A worked example
Imagine four observed paths, two of which converted. Removing Remarketing drops the conversion probability to zero (every winning path went through it), while removing Facebook only reduces it to one in nine. After normalising the removal effects across the channels, the credit might land at roughly Remarketing 41%, Facebook 32%, and Google 27% of the two conversions. Notice how the model rewards the channel the journeys genuinely depended on, rather than whichever happened to be first or last.
Strengths and weaknesses
Markov chains are a workhorse behind a lot of data-driven attribution, and for good reason:
- They cope better with smaller datasets and more granular channel splits than heavier methods like Shapley value attribution.
- They need less computing power, and the results tend to align with linear models as a sanity check.
- They are less rattled by statistically insignificant random data.
Here is how Markov chains compare with the alternatives:
| Model | How credit is assigned | Best for | Watch out for |
|---|---|---|---|
| Markov chain | Removal effect: credit reflects how much conversion probability drops without each channel | Multi-touch journeys with granular channel data at moderate volume | Still correlational; can overweight longer paths |
| Shapley value | Averages each channel’s marginal contribution across every possible combination of channels | Rich journey data where channel interactions matter | Computationally heavy; needs approximation at scale |
| Last-click | 100% credit to the final touchpoint before conversion | Simple, high-volume transactional funnels | Ignores every touch that built the decision |
| Linear | Equal credit split across every touchpoint | Teams wanting a simple, unbiased baseline | Treats a passing glance and a decisive visit as equally important |
The weaknesses are real, though. The approach carries “an embedded error resulting from the use of the removal effect” that can unfairly shift credit toward longer paths. And the deeper caveat applies to every model of this type: “any algorithm that use conversion paths as input data is based on the analysis of correlations”, which means it can describe what tends to happen but cannot prove what caused it. For that, you need incrementality testing.
Questions to ask yourself
As a modern growth or agile marketing professional, ask yourself the following about Markov chain attribution:
- Do I have enough clean conversion-path data to make a probabilistic model meaningful?
- Am I treating the removal effect as a useful estimate, or mistaking it for proven causation?
- Does the model’s credit roughly agree with a simple linear model, and if not, why?
- Would a Markov model actually change a budget decision, or am I modelling for its own sake?
- Have I paired it with a holdout test to check the story against reality?
Related articles
- Marketing Attribution Models: The Definitive Guide
- Shapley value attribution
- What is data-driven attribution?
- What is multi-touch attribution?
- Incrementality
About Growth Method
Markov chain attribution is a sharper way to split credit, but it is still built on correlation: it can tell you how much a channel’s absence would have changed the numbers, not what actually happened if you switched it off. Growth Method is the agentic marketing platform for B2B teams, built around exactly that gap. Agents read your live Search Console, GA4 and PostHog data, turn a claim like “this channel deserves more of the budget” into a tracked campaign with a recorded baseline, and measure the result against your own first-party data rather than a model’s estimate.
Get started: first month free, no credit card required.
Frequently asked questions
What is Markov chain attribution?
Markov chain attribution is a data-driven model that maps customer journeys as a sequence of channel “states” and credits each channel using the removal effect: how much overall conversion probability drops when that channel is taken out of the graph.
How is the removal effect calculated?
The removal effect is calculated by building a graph of every observed conversion path, removing one channel at a time, and recalculating the probability of reaching a conversion. The percentage drop is that channel’s removal effect, and the effects across all channels are then normalised so they sum to 100% of credit.
Markov chain vs Shapley value attribution: what’s the difference?
Both are algorithmic, multi-touch models built on your actual conversion data rather than a fixed rule. Markov chain attribution measures how much conversion probability drops when a channel is removed from observed paths, while Shapley value attribution averages a channel’s marginal contribution across every possible combination of channels. Markov chains need less computing power and cope better with smaller datasets; Shapley value is more computationally expensive but grounded in game-theory fairness axioms.
Does Markov chain attribution prove which channel caused a conversion?
No. Like other path-based models, it is based on correlation in your conversion data rather than a controlled experiment. It can show how credit shifts under the model’s assumptions, but confirming true causal impact requires incrementality testing.