Innovation & Emerging Tech

Sustainable innovation in engineering organizations requires a structured process that separates exploration from exploitation — running the current business reliably while evaluating emerging technologies that could create future competitive advantage. The Three Horizons Model (McKinsey) categorizes work into H1 (core business optimization), H2 (adjacent opportunities, 1–3 years), and H3 (transformational bets, 3–5 years). ThoughtWorks Technology Radar (Adopt/Trial/Assess/Hold) provides a vocabulary and process for organizational technology evaluation. Time-boxed PoC frameworks with explicit success criteria prevent innovation theater where experiments run indefinitely without producing decisions.

Key Points

• ThoughtWorks Radar process: quarterly, a cross-functional council of senior engineers evaluates technologies into four rings; Adopt (use in production), Trial (low-risk experiments), Assess (research and prototype), Hold (avoid for new projects)
• PoC framework: define the hypothesis → timebox to 2 weeks → define pass/fail criteria before starting → dedicate one engineer full-time → demo to stakeholders → write a 1-page decision memo → adopt, defer, or discard
• Three Horizons: H1 = incremental improvements to current products (80% of engineering capacity); H2 = adjacent opportunities requiring new skills or markets (15%); H3 = transformational bets (5%); the ratio is a deliberate organizational choice
• Innovation anti-patterns: "innovation theater" (hackathons with no path to production), "shiny object syndrome" (adopting every new technology without evaluation), "not invented here" (refusing to adopt proven external solutions)
• Horizon scanning practices: reading committees (curated tech blogs, research papers, conference proceedings), attendance at KubeCon/QCon/re:Invent, internal tech talks (lightning talks from engineers on personal experiments), vendor briefings
• Emerging technology risk: evaluate on 5 dimensions: community maturity, production references at scale, security posture, operations tooling availability, and talent market depth (can you hire for it?)
• Innovation funding: dedicated "innovation time" (Google's famous 20% policy) or structured innovation sprints (one sprint per quarter per team for exploratory work); without protected time, innovation is perpetually deferred by delivery pressure
• Fast failure principle: the goal of an H3 bet is to fail fast and cheaply if the hypothesis is wrong, not to succeed slowly; a 2-week PoC that disproves a hypothesis is a success — it saved 18 months of misguided investment