Data Science in eCommerce — Part 4

Deep dive into customer journey analysis

In this part we will perform analysis to understand customer journey to conversion. We will uncover paths that brings the most valuable conversions. We will also analyse the impact of different marketing channels on the customer journey and conversion value.

So far we learnt two things: both conversion value and path length are not normally distributed. In fact both follows Pareto distribution. This will be our starting point for the analysis.

We will try to answer some questions:

  1. How conversion value changes depending on the length of customer journey?
Conversion value: Cumulative conversion value versus length of customer journey

Shorter customer journeys generated higher conversion value

2. How number of conversions changes depending on the length of customer journey?

Number of conversions: Cumulative number of conversions versus length of customer journey

Takeout: Shorter customer journeys generated more conversions

Above two charts can be nicely summarised by this diagram.

Note outliers on above diagram— check the previous part to learn how outliers skewed statistical summary of data. Visual analysis helps to identify outliers in the dataset.

3. Customer journeys deep dive

So far we uncovered the relationship between length of the customer journey and the value and the number of conversions. Couple of questions remain:

What is the actual volume of our customer journeys?

Distribution of the length of customer journey

Above diagram shows number of customer journey for each type of path length. Let’s zoom in:

Distribution of the length of customer journey

Takeout: Majority of the customers had a customer journey between 4 and 8 touchpoints.

4. Customer journeys deep dive — Grand finale
Let’s combine our findings together: map journey length histogram with impact on conversion value

Combining journey length histogram with impact on conversion value


  • Disparity between customer journeys that brings the highest conversion value (one touchpoint) and the most common customer journeys (five touchpoint).
  • Customer journey with single touch points generated the highest conversions value (40% of all conversion value). On other hand this is a small cohort of the overall number of customer journeys (diagram on the top).
  • The most common types of journey are made out of 5 and 6 touchpoints. Unfortunately, this type of paths yields relatively low conversion value (cumulative).

Above takeouts could be nicely illustrated be the graph below:

It shows cumulative conversion value (percentage) growing with each channel.

Some very important questions for the business:

  • Customer journey with one, two and three touchpoints are responsible together for almost 80% of all conversion value.
  • Customer journeys with up to 5 touchpoints generate 90% of all conversion value

What we can learn from the optimisation perspective?

  • Customers with more than 4/5 touchpoints are highly likely to do not generate any substantial revenue. Should there be a trigger to stop paid marketing activities for those users?
  • What is the channel makeup for the most successful journeys? Journeys with the short customer path — one, two three touchpoints?

We will find answers to the above two questions in the next part:

Part 5: Finding patterns

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.