Flavors of Analytics

Marty Cagan argues that product teams often fixate on a single kind of analytics, usually web analytics, and miss the larger picture. He lays out four flavors that matter for product work: user analytics on how real people use the product, customer analytics on usage across a customer's organization, business analytics that examine trends over time through a data warehouse, and custom compute analytics that instrument the product to test specific theories. Together they let teams see behavior at multiple levels. He closes by saying the product leader should be capable with all four and should own understanding of the product's data rather than delegating it.

01Key takeaways

• Do not rely on web analytics alone; combine user, customer, business, and compute analytics for the full picture.
• Use user analytics to see how individuals and groups interact with features, tutorials, and funnels.
• For business products, track usage and revenue per customer, not just per user, drawing on CRM and finance data.
• Build a data warehouse that joins sources so you can study metrics like lifetime value and churn across time.
• Instrument the product custom when you have a specific theory to test, such as correlating usage with customer value or finding performance bottlenecks.
• Stay personally capable with all these data types and take ultimate responsibility for understanding your product's data.

02Key sections

Why teams see only one kind of analytics

Many people default to web analytics and overlook the broader value of data. Combining analytics types gives a fuller view of the product.

User analytics

This flavor tracks how individuals and aggregate groups actually use the app or service, such as tutorial usage, feature access, and funnel progress. Google or Adobe tools are common choices.

Customer analytics

For business products with many users per account, this view looks at usage, support load, time to value, and revenue across each customer, often drawing from CRM and financial systems.

Business analytics and compute analytics

Business analytics uses a data warehouse to study lifetime value, churn, and acquisition cost over time. Compute analytics involves custom instrumentation to test specific hypotheses, such as which operations are slow.

Responsibility of the product leader

Product leaders need enough skill and access to use all four types and should own their product's data, even while relying on analysts and scientists for support.

03From the post

“A partnership dedicated to teaching best practices to product teams and product leaders”

04Frameworks mentioned

Summary and takeaways written by PM Atlas; quotes are short excerpts. © the original author.