How You Battle the Data Wheel of Death in Growth

Summary

Balfour identifies the “Data Wheel of Death” — a vicious cycle where data isn’t maintained, becomes irrelevant, people lose trust, they use it less, and maintenance drops further. The root cause is treating data as a project instead of a product. Core mental models:

Project Mindset vs. Process Mindset. Data is never done. Four reasons: (a) your product will change, (b) your understanding of the business will change, (c) new answers expose new questions, (d) shit breaks. Andrew Chen: “Your data and KPIs should be a reflection of your strategy.” Strategy changes; data must follow.
The Data Team as Product Team. The data team’s customers are internal users. They need to define customer segments, understand needs, deliver compelling solutions, and iterate. Their output must enable other teams’ output, not make the data team the exclusive bottleneck.
Reward Systems Drive Behavior. Teams do what they’re rewarded for. Four reward types: financial (bonuses/raises/equity), progression (promotions), authoritative recognition (boss praise), peer recognition. If you want data-driven behavior, make it part of all four reward types. When you promote someone, announce the specific behaviors that led to it.
KPI Ownership Triad. Setting a KPI isn’t enough. Three things must be true: (a) the team feels ownership over it, (b) every person understands it and has easy access to view it, (c) the KPI is part of the reward system but not the only thing rewarded.
Hire Business Savvy, Not Just Technical. Don’t just ask “how do we calculate this metric?” Ask candidates “in this business scenario, what metrics do you think would be important?” The interview question reveals whether someone is a tool operator or a business thinker.

Relevance

06-reference/2026-04-03-reforge-why-analytics-efforts-fail — The Wheel of Death is the mechanism by which analytics efforts fail. Balfour’s process mindset is the prevention.
06-reference/2026-04-03-data-maturity-processes-tools — Data maturity requires breaking the Wheel of Death at each stage. The project-to-process shift is the maturity inflection point.
06-reference/2026-04-03-good-data-scientist-bad-data-scientist — “Good DS” traits map to Balfour’s hiring advice: business-savvy, proactive, ownership-oriented.

Open Questions

How do you measure whether your data team is operating as a product team vs. a service desk? What are the leading indicators?
At what company stage should you hire a dedicated “data culture” promoter vs. expecting it to emerge organically?