Standard : Hypothesis-driven development is practiced across teams

Purpose and Strategic Importance

This standard promotes the use of hypothesis-driven development (HDD) to reduce risk and increase learning when building digital products. By explicitly stating assumptions and testing them through measurable experiments, teams ensure that what they deliver actually drives the intended outcomes.

It supports the policy “Shape Roadmaps to Learn, Not Just to Deliver” by encouraging delivery that validates value incrementally rather than assuming success up front. HDD allows teams to pivot early, double down on what works, and stop investing in features that don’t deliver value.

Strategic Impact

  • Reduces wasted effort by validating assumptions early
  • Aligns delivery with business learning goals, not just feature output
  • Encourages continuous feedback and iterative discovery
  • Enables faster course correction and improved product-market fit
  • Fosters a culture of experimentation, accountability, and innovation

Risks of Not Having This Standard

  • Teams deliver features without clear understanding of their impact
  • Product decisions are based on assumptions, not evidence
  • Time and money are spent on initiatives that don’t solve real problems
  • Roadmaps become rigid and speculative, slowing adaptation
  • Learning opportunities are missed, undermining value delivery

CMMI Maturity Model

Level 1 – Initial

Category Description
People & Culture - Delivery is focused on output; assumptions go unchallenged.
Process & Governance - Features are planned and shipped without testing hypotheses.
Technology & Tools - No support for instrumentation or measurement of outcomes.
Measurement & Metrics - No validation of value delivered; success is binary (shipped or not).

Level 2 – Managed

Category Description
People & Culture - Some teams discuss intended outcomes informally.
Process & Governance - Hypotheses are occasionally included in discovery work.
Technology & Tools - Basic analytics or A/B testing tools available but underused.
Measurement & Metrics - Success metrics are defined inconsistently across teams.

Level 3 – Defined

Category Description
People & Culture - Teams routinely define hypotheses and design experiments.
- Failure to validate is seen as learning, not failure.
Process & Governance - HDD practices are embedded in delivery lifecycle.
- Roadmaps include testable outcomes, not just features.
Technology & Tools - Feature flags, analytics, and experimentation platforms are in use.
Measurement & Metrics - % of features with stated hypotheses and success criteria.

Level 4 – Quantitatively Managed

Category Description
People & Culture - Teams prioritise work based on validated learning potential.
Process & Governance - Experiments are tracked, outcomes reviewed, and learnings disseminated.
Technology & Tools - Automated A/B testing, cohort analysis, and feedback loops in place.
Measurement & Metrics - Learning velocity; % of roadmap items driven by validated hypotheses.

Level 5 – Optimising

Category Description
People & Culture - Hypothesis framing is a core part of team culture.
- Teams challenge assumptions proactively and collaboratively.
Process & Governance - HDD shapes portfolio decisions and investment strategies.
Technology & Tools - Predictive models inform which hypotheses to pursue.
Measurement & Metrics - Success is defined in terms of learning outcomes and value realised.

Key Measures

  • % of roadmap items framed as testable hypotheses
  • Number of experiments run and evaluated per quarter
  • Time to validated learning per feature
  • Ratio of validated to invalidated hypotheses
  • Business outcomes influenced by experiment-driven iterations