Scaling Data: Data Informed to Data Driven to Data Led
Summary
Reforge introduces a three-stage data maturity framework -- Data Informed, Data Driven, Data Led -- and argues that the sequence Strategy > Stage > Team > Tools is the correct order of operations for scaling data capabilities. Most organizations make the mistake of starting with tools or team hires, when data needs to be seen as a strategic lever for growth first.
The three stages:
- Data Informed -- focused on product-market fit. Business needs: operational visibility, KPI dashboards, metric definitions. Pitfall: allowing multiple versions of truth to coexist.
- Data Driven -- feature-level optimization. Business needs: self-serve analytics, experimentation tooling, deep-dive insights answering why not just what. Pitfall: misaligned incentives between data teams and the rest of the org.
- Data Led -- the business cannot operationally function without data products. Business needs: automated decision-making, near-real-time data, feature engineering. Pitfall: over-engineering infrastructure ahead of actual need.
Three recurring analytic problems thread through all stages:
- Misalignment between product strategy and available data capabilities
- Stage mismatch -- hiring people whose experience is from a different maturity level than the company
- Incorrect incentives -- measuring data teams on answers delivered rather than business impact
This framework is extremely useful for [[01-projects/phdata/index]] and [[01-projects/phdata/career-transition]]. In consulting, the first question is always "where is this client on the maturity curve?" -- the answer determines what to recommend. Recommending a data mesh to a Stage 1 company is malpractice; recommending manual CSV exports to a Stage 3 company is negligence.
The "Strategy > Stage > Team > Tools" hierarchy reinforces [[06-reference/2026-04-03-data-maturity-processes-tools]] -- both sources agree that tool selection should follow from decision-making needs, not lead them. It also connects to [[06-reference/2026-04-03-feature-stores-hierarchy]]: feature stores are a Data Led capability, not a Data Informed one.
For [[01-projects/data-marketplace/index]], this framework helps position the offering. A data marketplace is a Data Driven/Data Led product -- it assumes consumers already have operational visibility and are ready to use external data for optimization or automation.
The incentive misalignment problem is a version of the org design challenge in [[06-reference/2026-03-31-block-hierarchy-to-intelligence]] -- if the hierarchy rewards question-answering over impact, the data team will optimize for throughput rather than leverage.
Open Questions
- How do you diagnose which stage a company is actually at vs. which stage they think they are at? Are there specific signals beyond the ones listed?
- The stage mismatch problem applies to consultants too -- how do you ensure a consultant's playbook matches the client's maturity?
- What does "Data Led" look like for a company of 10-50 people, or is it exclusively a large-company phenomenon?