(Optional) Further Reading (Optional) Further Reading If you are interested in exploring topics we covered in this week’s content further, here are some good resources to continue your foray into product metrics.

Growth Accounting

We touched on the concept of the ‘Total Addressable Market’ (TAM), but haven’t really discussed ways to measure the TAM penetration, the growth stage of a product or product-market fit. This blog post from Sequoia Capital discusses approaches to tackle this. Other articles in their Data-Informed Product Building series are great to refresh other concepts we discussed this week.

Also, while discussing retention and engagement, we didn’t cover ways to further segment users into engagement tiers. Not every user engages with your product every day. Sometimes they may skip a day (go ‘dormant’) and then come back (become ‘resurrected’). Duolingo shared their experience using these concepts to create a quantitative growth model for their product and to identify growth opportunities.

Causal Inference

It can be hard to identify if a given relationship between two events’ occurrences is truly a cause-and-effect relationship, or just a correlation (and there is a third event or trait, a confounder, that makes these two events co-occur). If your customer accesses your product on a mobile app in addition to the website, and also happens to have an above-average spend with your product, is that a cause-and-effect relationship? Subsequently, does this mean we should drive all users to the mobile app? Or are both of these things just a manifestation that they are a high-intent user? To figure this out, we could construct an experiment (more on that in Week 4!), or use a statistical approach called causal inference that helps us construct models that isolate the impact of possible confounders.

Uplimit has a great course on causal inference! For those who want a more asynchronous learning path or may be less technical, a book called “Mastering ‘Metrics” is a good conceptual foray into causal inference. For more technical users, Causal Inference in Python and Causal Inference The Mixtape are great resources.

Cohort Analysis

We covered the basics of cohort analysis for product metrics quite a bit this week. If you want more visualization examples and cohort analysis applications to revenue metrics (like Revenue Retention, Monthly Recurring Revenue), this is a great and comprehensive blog post.