The innovation portfolio review that replaces vanity metrics

Why activity metrics fail

Hackathons, pilots and partnership counts can prove that a team is busy without showing whether the company is becoming more capable. The review should begin with the strategic uncertainty each initiative is meant to reduce and the decision that new evidence could unlock.

The OECD's measurement guidance distinguishes innovation activities from innovation outcomes. That distinction is useful inside a company: resources and experiments describe effort, while adoption, capability and business effects describe progress.

Review the portfolio through four lenses

A balanced review asks whether the portfolio covers the right strategic domains, whether initiatives are producing credible evidence, whether the organization can absorb what works and whether the total exposure still fits the company's risk appetite.

• Strategic fit: which future advantage or vulnerability does the initiative address?
• Evidence velocity: what uncertainty was reduced since the previous review?
• Adoption readiness: which operating owner, data and process will carry the result?
• Option value: what can the company do next that it could not do before?

Create explicit continuation rules

Every initiative should enter the review with a next decision already defined: expand, redesign, hold, transfer or stop. This prevents teams from treating more experimentation as the default answer when evidence is weak.

A strong portfolio is not one where every project succeeds. It is one where the company learns early, moves resources deliberately and turns the few credible signals into operating advantage.

Build a common evidence scale

Portfolio comparison fails when one team reports technical milestones, another reports customer enthusiasm and a third reports revenue without showing what uncertainty each result reduced. A common evidence scale can classify claims as assumptions, observations, repeated results and operating proof. The label refers to the specific claim, not the prestige of the project or team.

Evidence quality also depends on relevance. A controlled laboratory benchmark can be rigorous but weak evidence for production economics; a paid pilot can indicate demand but say little about retention or scale. Reviewers should ask which decision the evidence supports and which important uncertainty remains untouched.

• Claim being tested
• Method and comparison baseline
• Representativeness of conditions
• Decision unlocked by the result

Run the review as a resource market

Projects should compete for scarce engineering, executive and integration capacity, not only a budget line. The review shows resources already committed, the next smallest evidence-producing step and the opportunity displaced if work continues. This makes hidden congestion visible and rewards projects that can learn economically.

Stopping is accompanied by a knowledge-transfer plan: evidence, reusable components, customer learning, failure conditions and people who carry the insight. A stopped initiative can strengthen the portfolio if it frees capacity early and prevents the same assumption from being retested without context.

• Next decision and evidence threshold
• Incremental resource request
• Operating owner for transfer
• Archive and reuse of learning