Metric trees are a framework for modeling underlying business processes, varying in complexity based on the interconnectedness of process steps and their evolution over time.
Using a classic 2x2 visual, let’s use the standard eCommerce business model to illustrate these varying possibilities.
1. Process is “Tightly Connected” + Time is “Aligned”:
In eCommerce, the process of users adding items to their carts and checking out tightly couples the steps in the process and aligns in time with the completion of a transaction.
The transaction basket here can be decomposed into an arithmetic tree containing items per order, units per item, price per unit, and discounts per unit if any.
2. Process is “Loosely Connected” + Time is “Aligned”:
In revenue tracking, growth and finance teams aggregate transactions across different user segments over fixed time periods (e.g., monthly or quarterly). Although transactions align in time, the underlying processes of how new and repeat customers transact are only loosely related.
Decomposing new or repeating customers into their respective marketing channels for example, further illustrates the “loosely connected” nature of the underlying metrics.
3. Process is “Tightly Connected” + Time is “Offset”:
Consider the scenario where these aforementioned transactions can be modified before they are shipped. In addition, consider that the items can ship at different times, and after they are shipped, the user could return one or more items and request credits or refunds. Once these items are mailed back, they possess varying degrees of reusability and net salvage value.
If we were to construct a metric tree that decomposes a profit equation into its component parts, we are dealing with a tightly connected process that is still tied to the original transaction and the associated basket of items but one that involves various steps that unfold over time.
4. Process is “Loosely Connected” + Time is “Offset”:
For retention metrics in eCommerce, tracking how users progress from their initial to subsequent transactions spans longer time frames and involves loosely connected processes. Metric trees can model these user journeys over time, focusing on how retention rates vary across different user cohorts.
When designing metric trees for your business model, consider these axes — the process and time grains —to structure relationships effectively and to capture the necessary intricacies.
