Data-Informed Product Management

In doing Agile Product Development, there can be a tension between a "top-down" (I don't mean exec-driven, I mean mental-stack) Product Vision/Product Strategy/Conceptual Integrity and an incremental/iterative process of Thinking In Bets.

waterfall: spend years building the ultimate product before launching it
incremental waterfall: deliver piecemeal, but filter every lack of adoption as "we just need to get further along the roadmap"

A new anti-pattern: being mindlessly data-driven

"since FeatureAPage1 gets more views than FeatureBPage1, FeatureA is more important/valued"
"we used a survey to have users rank the importance/value of 6 features - we must put the menu items in that order"
"people's biggest request is FeatureX so we must do that"

Solution: A smarter way to use data

see Quadruple-Loop Product Management
iterate on your mission, business model (lean canvas), product vision, product strategy: these are all models (see Strategic Context)
- identify/rank the Risks (see Lean Stack)
- continuous discovery
identify/rank problems/opportunities
identify/rank initiatives/solutions -> Opportunity Solution Tree
recognize that each initiative-candidate is a Bet - generate ideas on how to "test" that Bet with as little work/delay as possible
- user phone interviews, user email conversations, user online surveys, prototyping for User Research.... MVF/MVP
released work generates adoption Data which with thinking supports/challenges your models
- Doing SQL Data Analysis As A PM
if it seems to challenge/kill your plan, consider where to change direction (micro-pivot)
- feature design
- solution ranking
- problem validation
- strategy validation
- vision validation

Bill Seitz