Customer Journey & Attribution

The path from first impression to revenue.

Every customer interaction stitched together, visualised, then credited five different ways — including a real Markov chain model running on these journeys.

Multi-touch journey flow

First touch → middle touchpoint → last touch → outcome. Channels below 2% share are collapsed into "Other" for legibility.

Avg path length

1.2

touchpoints per conversion

First touch

Middle touch

Last touch

Outcome

Attribution model comparison

Same conversions, credited five different ways. The Markov row is the result of an absorbing-Markov-chain removal-effect calculation on this dataset — not a guess.

Path length distribution

How many touchpoints did each customer have? Most customers cluster around 1 touchpoints.

Average time-to-conversion by first-touch channel

From the very first impression to the purchase event. Channels with short windows tend to attract high-intent prospects; long windows indicate top-of-funnel awareness.

Google Ads

2.6d

median · mean 9.7d

Direct

2.7d

median · mean 11.2d

Meta Ads

2.8d

median · mean 18.8d

Referral

2.8d

median · mean 18.6d

Organic

2.8d

median · mean 13.3d

Snapchat

2.9d

median · mean 22.0d

3.5d

median · mean 21.2d

TikTok

3.6d

median · mean 2.4mo

How to read this

Last-click is the model platforms report by default. It gives 100% credit to whichever channel was last seen — and is structurally biased toward bottom-funnel channels like Google Search.

First-click swings to the opposite extreme, crediting the discovery channel and ignoring everything that happened in between.

Linear & time-decay split credit across the path with mechanical weighting rules.

Markov chain is what's actually running here: I build an absorbing-Markov-chain transition matrix over every customer journey, then compute the removal effect — how much would conversions drop if a given channel disappeared? — and normalize to 100%. The number under each channel in the bar above answers: how important is this channel to the chain reaching a conversion?

Full methodology, including the math and code, is on the /docs page.

Total Markov-credited revenue weight across channels: 100%