â–¶Why do most corporate innovation labs fail within 2-3 years?
Three root causes: (1) Organizational misalignment—labs report to business units with operational KPIs, so innovation projects die when quarterly targets slip. Fix: separate governance, 24-36 month patience cycles. (2) Wrong metrics—labs measure outputs (ideas generated, patents filed) instead of learning velocity (hypotheses tested, pivot speed). Fix: use innovation accounting (minimum viable features, validated learning). (3) Insufficient protection—innovation teams get pulled into firefighting or forced to adopt waterfall processes designed for scale, not exploration. Fix: physical/cultural separation + explicit anti-bureaucracy mandate from executives.
â–¶What's intrapreneurship and how do I build a culture that supports it?
Intrapreneurship = employees starting new ventures inside the company. Google's 20% Time and Amazon's Day One Fund are examples. Culture requirements: (1) Psychological safety (failures don't tank careers), (2) Access to capital without 5 layers of approval, (3) Clear exit criteria (when does a project get spun out vs killed vs handed to ops?), (4) Founder rewards (equity, promotions, autonomy for successful pivots). Common mistake: incentivizing idea quantity over idea quality; 100 bad ideas ≠innovation culture.
â–¶How do I apply McKinsey's 3 Horizons to my portfolio?
Horizon 1 (70% budget/effort): core business, incremental improvements, 1-3 year payoff. Horizon 2 (20%): adjacent markets, new customer segments, existing capabilities, 3-5 years. Horizon 3 (10%): transformational, moonshots, may cannibalize H1, 5-10+ years. The math: if you only invest H1, competitors eat your lunch. If you over-invest H3, you starve operations. Lock in the 70/20/10 ratio via quarterly reviews—if H3 drops below 8%, you're drifting toward commodity. Case: Kodak had great innovation labs but allocated <2% to digital; missed the paradigm shift.
â–¶What's the difference between innovation ventures and corporate venturing?
Innovation ventures = new products/services launched inside the company (e.g. Netflix moving from DVDs to streaming). Corporate venturing = companies investing in or acquiring external startups to learn, acquire talent, or diversify. Innovation ventures are cheaper to scale (internal ops, brand, distribution) but require culture shift. Corporate venturing is faster to capability but risks integration hell (cultural clash, talent exodus post-acquisition). Best practice: do both—allocate 15-20% to ventures inside, 5% to strategic investments outside.
â–¶How do I measure innovation ROI when projects have 5-10 year payoffs?
Traditional ROI (profit ÷ cost) fails for innovation; use innovation accounting instead: (1) Track learning velocity (hypotheses → validated learning per month), (2) Option value (cost to keep option alive = cost of next experiment; cheaper than full build), (3) Pivot speed (time from 'wrong hypothesis' to 'new direction'), (4) Ecosystem health (partner interest, talent attraction). Quantify: project costs $500k/year; if it validates 5 critical assumptions in year 1, option value = $5M (what a venture studio would pay for that learning).
▶AI as innovation accelerator 2026—how do I use it without hype?
Real use cases: (1) Rapid prototyping (AI generates 50 wireframes, humans pick 5, iterate), (2) Market research automation (scrape competitor features, sentiment analysis on 10k reviews in hours), (3) Hypothesis generation (AI flags non-obvious adjacencies in customer data), (4) Testing velocity (A/B test variants 5x faster with AI design). Hype: 'AI will innovate for us' / 'AI replaces ideation'. Truth: AI is a research assistant + tooling accelerator; humans still pick direction and kill bad ideas. Budget: 20% of lab budget → AI tools + training; don't chase every OpenAI launch.
â–¶How do I balance protecting innovation teams with keeping them connected to business reality?
Common mistake: innovation labs become ivory towers, disconnected from actual customer/operator needs. Fix: (1) Quarterly reviews with ops leaders (not approval, but input), (2) Mandatory customer immersion (labs → talk to 10 users/month), (3) Clear handoff criteria ('when this project exits the lab, ops owns it'), (4) Failure retrospectives public (not punitive—normalize learning failures). Red flag: if ops teams say 'we don't know what the lab is doing', the lab has drifted. Bridge: embed ops liaison in lab quarterly planning.