Markov Chain Attribution Explained: Markov chain attribution credits channels by the removal effect — the drop in conversion probability when a channel is deleted from the journey graph. Here is how it works, where it quietly misleads DTC brands, and why it still cannot prove incrementality.
Read the full article below for detailed insights and actionable strategies.
The attribution problem
One sale. Four channels. 400% credit claimed.
Reported revenue: €400 · Actual revenue: €100 · Gap: €300
The 60-Second Answer
Markov chain attribution is a data-driven model that treats each customer journey as a sequence of channel "states" and credits every channel by its removal effect — how much total conversion probability drops when that channel is deleted from the graph. It distributes credit more fairly than last-click, but it is still correlational: it explains observed paths, not incremental causation.
What Markov Chain Attribution Actually Is
Markov chain attribution borrows a tool from probability theory. It models the buyer journey as a series of transitions between states — where each state is a marketing channel (Paid Search, Email, Organic, Direct) plus two special states: Start and the absorbing states Conversion and Null (no conversion).
The model's defining assumption is the Markov property: the probability of the next step depends only on the current state, not the full history that preceded it. A first-order Markov chain "remembers" only where the customer is right now, not the winding route they took to get there.
From thousands of real journeys, the model builds a transition matrix — the probability of moving from any channel to any other, or to Conversion. Then it computes credit using the removal effect: remove one channel from the graph, recalculate the overall conversion probability, and measure the drop. The bigger the drop, the more essential that channel was to the network of paths. This makes it a form of algorithmic attribution and a close cousin of Shapley value attribution.
In one sentence: Markov credits a channel by asking "how much worse would conversions be if this channel vanished from every path we observed?"
How the Removal Effect Works (Step by Step)
Here is the workflow a Markov model runs, and the one you would replicate if you built it yourself in Python or R's ChannelAttribution package.
- Collect path data. Export every converting and non-converting journey as an ordered sequence of channels. In practice this comes from a GA4 export or a data warehouse.
- Normalize the channel names. "facebook", "Facebook", and "FB / paid" must collapse into one state. Inconsistent UTM parameters silently fracture a channel and corrupt every downstream number.
- Build the transition matrix. Count how often each channel leads to each next state, then convert counts to probabilities.
- Compute the baseline conversion probability of the whole graph from Start to Conversion.
- Remove one channel (route its traffic to Null) and recompute the conversion probability.
- Measure the drop. The percentage decrease is that channel's raw removal effect.
- Repeat for every channel, then normalize so all credits sum to 100% — because raw removal effects do not naturally add up to one.
- Allocate revenue to each channel in proportion to its normalized removal effect.
The output looks reasonable, scientific, and far better than last-click attribution. That is exactly why it is dangerous to over-trust.
The Markov Trust Ladder (Original Framework)
A Markov result is only as trustworthy as the data and the decision behind it. Use this four-rung ladder — built for DTC operators, not statisticians — to decide how much weight to put on the numbers before you move budget.
| Rung | What you have | How much to trust Markov | What to do |
|---|---|---|---|
| 1 — Decorative | < 300 monthly conversions, messy UTMs | Almost none | Treat output as a conversation starter, not a budget input |
| 2 — Directional | 300–1,000 conversions, clean tagging | Low–medium | Use to rank channels, never to set exact spend |
| 3 — Diagnostic | 1,000+ conversions, deduplicated cross-device paths | Medium | Use to flag channels worth an incrementality test |
| 4 — Decision-grade | High volume plus a validating holdout or geo test | Medium–high | Use for budget — but only because causation was checked separately |
The uncomfortable truth: most Shopify and DTC brands sit on rungs 1–2 and act as if they are on rung 4. Markov gives a precise-looking percentage to every channel, and precision is routinely mistaken for accuracy. The ladder exists to break that illusion.
Markov vs Other Attribution Models
| Model | How it assigns credit | Strength | Core weakness |
|---|---|---|---|
| Last-click | 100% to final touch | Simple, auditable | Ignores everything upstream |
| Linear | Equal split | Fair-feeling | Treats a banner view like a branded search |
| Time-decay | More to recent touches | Reflects recency | Arbitrary half-life |
| Position-based | 40/20/40 | Honors first + last | Rules are guesses |
| Shapley value | Game-theory marginal contribution | Theoretically fair | Computationally heavy; still observational |
| Markov chain | Removal effect across the path graph | Captures mid-funnel, data-driven | Correlational, volume-hungry, path-dependent |
| Causal / incrementality | Measured lift vs a control | Answers "did it cause sales?" | Needs a valid counterfactual |
Markov sits with data-driven attribution and machine-learning attribution: a major upgrade over rules-based models, but living on the wrong side of the correlation-versus-causation line.
Where Markov Quietly Misleads DTC Brands
1. It is correlational, not causal. The removal effect describes how channels co-occur on paths to purchase. It cannot tell you whether a channel caused incremental revenue or simply appeared on journeys that would have converted anyway. This is the same trap that makes multi-touch attribution fail: observing a touchpoint is not proving its lift.
2. It cannot see demand it did not capture. If a customer saw a YouTube ad, never clicked, and bought a week later via Direct, Markov often credits Direct — because the un-clicked impression rarely enters the path data. Channels measured only by view-through conversions are systematically distorted, just as they are in view-through ROAS.
3. It rewards channels that platforms already over-claim. Feed Markov path data assembled from platform pixels and you inherit self-attribution bias — Meta, Google, and TikTok each insisting they were on the path.
4. It cannot distinguish incremental from cannibalized sales. Brand search and retargeting harvest demand other channels created. Markov happily hands them credit, exactly the cannibalization blind spot every path-based model shares.
5. It is hungry and fragile. Below ~1,000 conversions the transition probabilities are mostly noise, and short attribution windows truncate the very journeys the model exists to analyze.
A €-Denominated Worked Example
A Shopify skincare brand spends €40,000/month and records 2,000 conversions. Their Markov model returns:
| Channel | Markov-credited revenue | Implied verdict |
|---|---|---|
| Branded Search | €78,000 | "Scale it" |
| Meta Prospecting | €52,000 | "Holding steady" |
| Retargeting | €46,000 | "Efficient" |
| €24,000 | "Fine" |
Branded search tops the table, so the team shifts another €6,000/month into it. Three months later, total revenue is flat.
Why? Branded search mostly harvests demand created upstream by Meta prospecting and email. Its high removal effect reflects that almost every path passes through it near the finish line — a correlational artifact, not incremental value.
A causal attribution read — validated with a branded-search holdout — tells a different story:
| Channel | Markov says | Causal lift says | Real action |
|---|---|---|---|
| Branded Search | €78,000 | €19,000 incremental | Cap spend; demand exists already |
| Meta Prospecting | €52,000 | €71,000 incremental | Underfunded — scale this |
| Retargeting | €46,000 | €12,000 incremental | Trim; mostly cannibalization |
| €24,000 | €28,000 incremental | Healthy; protect it |
Same data, opposite budget decision. Markov ranked the harvester first; causal inference ranked the creator of demand first. (Figures are illustrative.)
Common Mistakes
- Treating removal-effect percentages as incremental lift. They are not the same quantity.
- Running Markov on too little data and reading three-decimal precision into noise.
- Skipping channel normalization, so one channel is split across three misspelled states.
- Ignoring un-clicked impressions, then wondering why upper-funnel channels look weak.
- Never validating against a holdout or geo test. Markov should generate hypotheses; experiments confirm them.
- Using a 7-day window on a brand with a 30-day consideration cycle.
Checklist: Before You Trust a Markov Result
- At least ~1,000 conversions in the analysis window
- Channel names normalized and UTMs audited
- Cross-device paths stitched, not double-counted
- Attribution window matches your real buying cycle
- View-through impressions accounted for, not silently dropped
- Top channels flagged for an incrementality test
- Result cross-checked against new-customer CAC
- You know which Trust Ladder rung you are on (see framework above)
Key Takeaways
- Markov chain attribution credits channels by removal effect — the drop in conversion probability when a channel is deleted from the path graph.
- It is a genuine upgrade over rules-based models and a sibling of Shapley and data-driven attribution.
- It remains correlational: it explains observed customer journeys, it does not prove incrementality.
- It is data-hungry, sensitive to tagging, blind to un-clicked demand, and prone to crowning demand-harvesters.
- Use Markov to rank and hypothesize; use causal inference and incrementality testing to decide. See also the unified measurement framework and MMM vs MTA vs causal inference.
For €99, upload any historical GA4 period and get causal attribution for every channel in 5–10 minutes — no pixel, no migration. Go Pro at €299/mo for continuous attribution, an AI chatbot for your data, and a developer API.
Get attribution insights in your inbox
One email per week. No spam. Unsubscribe anytime.
Key Terms in This Article
Algorithmic Attribution
Algorithmic Attribution is a data-driven model using machine learning to analyze each touchpoint's impact in the customer journey. It assigns conversion credit by statistically evaluating both converting and non-converting paths.
Attribution Model
An Attribution Model defines how credit for conversions is assigned to marketing touchpoints. It dictates how marketing channels receive credit for sales.
Attribution Window
Attribution Window is the defined period after a user interacts with a marketing touchpoint, during which a conversion can be credited to that ad. It sets the timeframe for assigning conversion credit.
Causal Attribution
Causal Attribution uses causal inference to determine which marketing touchpoints genuinely cause conversions, not just correlate with them.
Causal Inference
Causal Inference determines the independent, actual effect of a phenomenon within a system, identifying true cause-and-effect relationships.
Customer journey
Customer journey is the path and sequence of interactions customers have with a website. Customers use multiple devices and channels, making a consistent experience crucial.
Incrementality Testing
Incrementality Testing measures the additional impact of a marketing campaign. It compares exposed and control groups to determine causal effect.
Multi-Touch Attribution
Multi-Touch Attribution assigns credit to multiple marketing touchpoints across the customer journey. It provides a comprehensive view of channel impact on conversions.
Related Articles
Sixty-second versions of these ideas: Causality Engine on YouTube Shorts.
Ready to see your real numbers?
Own the budget? Upload your GA4 export and see which channels drive incremental sales, with confidence intervals, in minutes. Have to defend it? Start with the live demo and take the read to your CFO.
Full refund if you don't see value.
Stay ahead of the attribution curve
Weekly insights on marketing attribution, incrementality testing, and data-driven growth. Written for the person who owns the budget and the person who has to defend it.
No spam. Unsubscribe anytime. We respect your data.
Frequently Asked Questions
What is Markov chain attribution in simple terms?
It is a data-driven attribution model that maps customer journeys as sequences of channel states and credits each channel by its removal effect — the drop in overall conversion probability that occurs when that channel is removed from the path graph. Channels whose absence hurts conversions most receive the most credit.
What is the removal effect?
The removal effect measures how much a channel matters by deleting it from the Markov graph, routing its traffic to a non-conversion state, and recalculating the probability of converting. The percentage drop is that channel's importance. Because raw removal effects do not sum to 100%, they are normalized before credit is assigned.
How is Markov attribution different from Shapley value attribution?
Both are algorithmic, data-driven models that improve on rules-based attribution. Shapley borrows from cooperative game theory and averages each channel's marginal contribution across coalitions, while Markov uses path-graph transition probabilities and the removal effect. In practice they often rank channels similarly, and both remain correlational rather than causal.
Is Markov chain attribution causal?
No. Markov attribution explains patterns in the journeys you observed; it does not establish that a channel caused incremental sales. A channel can earn a high removal effect simply by appearing on many paths that would have converted anyway. To measure true causation you need incrementality testing or causal inference.
How much data do you need for Markov attribution to be reliable?
As a rule of thumb, you want at least roughly 1,000 conversions in the analysis window before the transition probabilities are stable. Below a few hundred conversions the model mostly reflects noise, and the precise-looking percentages it returns should not drive budget decisions.
Why does Markov attribution over-credit branded search and retargeting?
These channels usually sit near the end of journeys and appear on almost every converting path, so removing them causes a large probability drop and a high removal effect. But they typically harvest demand created upstream rather than generating it, so their incremental value is far lower than their Markov credit suggests.
Should I use Markov attribution or causal attribution?
Use Markov to rank channels and generate hypotheses about which deserve scrutiny. Use causal attribution and incrementality experiments to decide where to move budget. Causality Engine applies causal inference to your GA4 export so you measure incremental contribution rather than path co-occurrence.