This is a synthetic case aimed to illustrate how I would treat product growth challenges: from inquiry to execution.

I took a Big Query Google Merch Store public dataset - https://console.cloud.google.com/bigquery?p=bigquery-public-data&d=ga4_obfuscated_sample_ecommerce&t=events_20210131&page=table&project=project-3013a5d3-611f-4647-980 to explore.

The dataset has a limited timeframe & event scope + there’s no metadata that can be used for thorough segmentation, thus the data and conclusions are skewed by these limitations.

In the course of exploration I looked into the following inquiries:

  1. Which behavioral and user factors are most strongly associated with the likelihood of purchase?
  2. How long does it take from the first visit to a purchase, and where are we losing users over time?
  3. How does a user’s first-session experience influence their likelihood to return and generate revenue over time?

Inquiries & results

My takeaways:

Conclusion Subsequent actions/features/experiments
High-intent users who make at least 1 purchase, view 22,6 items vs. 1,1 items viewed on average by a non-purchaser. The # of items to be viewed could become a proxy metric for purchase probability. 1. Adding a banner carousel with promoted items/lists of items;
  1. Changing categories for specific items in ‘Popular on the Google Merch shop’ section;
  2. Whole catalog end-to-end navigation etc. | | The mode time to purchase is 6.8 minutes, while the median is 67, which informs us of the nature of a purchase: retail, fast, one that doesn’t require much thought & coordination. More time-based urgency can be employed. | 1. Discount for new buyers with a countdown for 1-3 hours;
  3. Brighter discount highlights;
  4. ‘Haven’t found what you’re looking for?’ window to collect merch ideas and other problems; | | The duration of a session of high engagement users with repeated purchases is 7.4 minutes compared to 37 seconds for low-engagement ones with more page views and scrolls. | 1. Triggering customer data collection in the first 8 minutes = more opportunities to retain visitors with high engagement potential;
  5. Research if the checkout is a purchase barrier and, if yes, simplify it to the maximum;
  6. More data should be collected from other sources: customer interviews, metadata etc. about the targeted purchaser profile to inform customer journeys and the presentation of content. |