---
# name: stable local skill identifier; must match the parent directory name.
name: three-horizons
# description: routing contract; names when to activate and when not to activate.
description: "Use when balancing an innovation, growth, transformation, or venture portfolio across McKinsey's Three Horizons: Horizon 1 current core businesses, Horizon 2 emerging growth businesses, and Horizon 3 future options. Covers concurrent portfolio balance, resource allocation (the 70-20-10 benchmark and its caveats), evidence maturity, governance and metrics by horizon, metered funding, ring-fenced budgets, transitions from option to emerging business to core, the collapsed-time critique that disruption can now arrive on core timelines, and the risk that short-term core demands or incentives starve future growth. Do NOT use for BCG growth-share portfolio allocation (use bcg-matrix), Ansoff product-market growth path selection (use ansoff-matrix), Blue Ocean value-curve redesign (use blue-ocean-strategy), scenario construction and stress-testing across alternative futures (use scenario-planning), OKR goal-setting (use okrs), or quantified probability-weighted valuation (use expected-value)."
# license: usage license for this skill.
license: MIT
# compatibility: runtime compatibility notes.
compatibility:
  notes: "Markdown, innovation portfolio reviews, corporate growth strategy, transformation roadmaps, venture portfolio planning, R&D portfolio reviews"
# allowed-tools: tools this skill expects to use when loaded.
allowed-tools: Read Grep WebSearch WebFetch
# metadata: Skill Metadata Protocol fields consumed by Skill Graph tooling.
metadata:
  # subject: primary browse shelf. One of twelve values: backend-engineering / frontend-engineering /
  # software-architecture / data-engineering / agent-ops / ai-engineering / quality-assurance /
  # design / reasoning-strategy / software-engineering-method / knowledge-organization / product-domain.
  subject: reasoning-strategy
  # scope: PRD-style free-text statement of what the skill teaches and excludes.
  scope: "Three Horizons innovation and growth portfolio analysis for companies, business units, product portfolios, R&D portfolios, transformation programs, and venture pipelines: classify initiatives as Horizon 1 current core, Horizon 2 emerging growth, or Horizon 3 future option; balance current performance with future opportunity creation; compare actual funding, talent, leadership attention, and decision rights with strategic ambition and risk appetite; match governance, metrics, evidence, and funding rhythm to each horizon; apply resource-allocation benchmarks (70-20-10) as a context heuristic rather than a law; account for the collapsed-time critique that H3 disruption can now be delivered on H1 timelines; layer an evidence-quality overlay on horizon labels; protect ring-fenced H2/H3 budgets; identify starvation, gap, speed, zombie-portfolio, transition, and innovation-theater risks; and convert the portfolio view into rebalance actions. Excludes BCG growth-share allocation, Ansoff product-market growth-path selection, Blue Ocean value-curve redesign, standalone OKR planning, scenario-planning alternative-future construction, quantified expected-value valuation, generic roadmap sequencing, and treating horizons as a fixed calendar backlog."
  # public: publishability / private-data gate. true = safe for public release; false = private/internal.
  public: true
  # taxonomy_domain: optional slash-delimited sub-path within the subject.
  taxonomy_domain: foundations/strategy
  # stability: lifecycle posture. experimental / stable / frozen / deprecated.
  stability: stable
  # keywords: fuzzy activation phrases. Keep at most 10.
  keywords:
    - three horizons
    - horizons of growth
    - innovation portfolio
    - growth portfolio
    - horizon 1
    - horizon 2
    - horizon 3
    - core adjacent transformational
    - future growth options
    - portfolio balance
  # triggers: exact-match activation phrases.
  triggers:
    - three-horizons
    - three-horizons-framework
    - horizons-of-growth
    - h1-h2-h3
    - innovation-portfolio
  # examples: realistic prompts that should activate this skill.
  examples:
    - "Use Three Horizons to review our innovation portfolio and show whether current core work is starving future growth."
    - "Use the Three Horizons framework to map these growth initiatives into Horizon 1 current core, Horizon 2 emerging growth, and Horizon 3 future options, then recommend rebalance actions."
    - "Build a Three Horizons view for our product roadmap: core optimization, emerging businesses, and future options."
    - "Review this R&D portfolio with Horizon 1, Horizon 2, and Horizon 3 governance and metrics."
    - "We have too many incremental bets and no future options. Use Three Horizons to diagnose the gap."
  # anti_examples: near-miss prompts that should route elsewhere.
  anti_examples:
    - "Classify our business units as stars, cash cows, question marks, or dogs by market growth and relative market share."
    - "Use Ansoff to choose between market penetration, market development, product development, and diversification."
    - "Use Blue Ocean Strategy to create a strategy canvas and ERRC grid for this category."
    - "Build four plausible future scenarios with signposts, no-regret moves, and contingent actions."
    - "Write quarterly OKRs for the innovation team."
    - "Use expected-value analysis to compare these investment options with probabilities, payoffs, and scenario values."
  # relations: typed graph edges to sibling skills.
  relations:
    suppresses:
      - skill: bcg-matrix
        reason: "three-horizons owns innovation and growth initiatives by maturity, uncertainty, governance, and transition path; bcg-matrix owns portfolio allocation across business units by market growth and relative market share"
      - skill: ansoff-matrix
        reason: "three-horizons owns portfolio balance across current core, emerging businesses, and future options; ansoff-matrix owns product-market growth path classification by existing/new products and markets"
      - skill: blue-ocean-strategy
        reason: "three-horizons owns how a portfolio balances current and future growth bets; blue-ocean-strategy owns value innovation and market-boundary reconstruction for a specific strategic move"
      - skill: scenario-planning
        reason: "three-horizons owns innovation portfolio balance across current core, emerging growth, and future options; scenario-planning owns constructing alternative external futures from critical uncertainties and stress-testing a strategy across them"
      - skill: okrs
        reason: "three-horizons owns portfolio diagnosis and rebalance across horizons; okrs owns objective and key-result setting for execution tracking"
      - skill: expected-value
        reason: "three-horizons owns qualitative portfolio maturity, governance, and evidence fit; expected-value owns probability-weighted valuation once outcomes and probabilities are modeled"
      - skill: playing-to-win
        reason: "three-horizons owns innovation portfolio balance after strategic direction exists; playing-to-win owns the integrated strategy choices that define where and how to win"
    related:
      - ansoff-matrix
      - bcg-matrix
      - blue-ocean-strategy
      - scenario-planning
      - playing-to-win
      - expected-value
      - okrs
      - swot-tows
      - value-chain-analysis
      - epistemic-grounding
      - methodology
    verify_with:
      - epistemic-grounding
      - methodology
      - expected-value
  # grounding: truth sources and failure modes for this skill.
  grounding:
    subject_matter: "Three Horizons as a portable innovation and growth portfolio framework"
    grounding_mode: universal
    truth_sources:
      - https://www.mckinsey.com/capabilities/strategy-and-corporate-finance/our-insights/enduring-ideas-the-three-horizons-of-growth
      - https://www.mckinsey.com/capabilities/growth-marketing-and-sales/our-insights/now-new-next-how-growth-champions-create-new-value
      - https://www.mckinsey.com/capabilities/strategy-and-corporate-finance/our-insights/investing-in-innovation-three-ways-to-do-more-with-less
      - https://www.mckinsey.com/capabilities/strategy-and-corporate-finance/our-insights/how-top-performers-use-innovation-to-grow-within-and-beyond-the-core
      - https://steveblank.com/2019/01/08/the-fatal-flaw-of-the-three-horizons-model/
      - https://hbr.org/2019/02/mckinseys-three-horizons-model-defined-innovation-for-years-heres-why-it-no-longer-applies
      - https://hbr.org/2012/05/managing-your-innovation-portfolio
      - https://www.strategyzer.com/library/how-companies-should-manage-their-innovation-portfolios
      - https://www.strategyzer.com/roi-for-innovation
      - https://www.bcg.com/publications/2024/innovation-systems-need-a-reboot
      - https://www.internationalfuturesforum.com/three-horizons
      - https://www.internationalfuturesforum.com/world-model-three-horizons
      - https://www.internationalfuturesforum.com/transformative-innovation
      - skills/reasoning-strategy/three-horizons/references/three-horizons-sources.md
      - skills/reasoning-strategy/three-horizons/references/upstream-displacement-2026-06-09.md
    failure_modes:
      - horizons_treated_as_sequential_calendar_phases
      - calendar_timings_attributed_to_the_original_framework_as_fixed_rules
      - horizon_three_disruption_speed_underestimated_treated_as_distant
      - horizon_one_cash_engine_neglected_or_demonized
      - horizon_two_starved_between_core_demands_and_future_theater
      - horizon_three_ideas_scaled_without_evidence_or_options_logic
      - horizon_three_treated_with_horizon_one_hurdle_rates_killing_exploration
      - same_metrics_and_governance_used_for_all_horizons
      - resource_split_applied_as_a_fixed_law_instead_of_a_context_heuristic
      - declared_strategy_not_reflected_in_funding_talent_or_attention
      - recognition_and_incentive_systems_reward_only_h1_outcomes
      - horizon_three_budgets_not_ring_fenced_so_starved_in_downturns
      - sequential_s_curve_assumed_when_reality_is_non_linear_or_step_change
      - portfolio_balance_claimed_without_resource_or_decision_evidence
      - innovation_theater_confused_with_real_option_creation
      - private_strategy_customer_or_financial_data_used_in_examples_or_evals
    evidence_priority: general_knowledge_first

  # mental_model: primitives and relationships in one paragraph.
  mental_model: "Three Horizons is a concurrent portfolio map for growth work. The primitives are a portfolio owner, strategic ambition, risk appetite, Horizon 1 current core businesses that generate today's value and cash flow, Horizon 2 emerging opportunities that may become material growth engines but need investment, Horizon 3 future options that are uncertain and exploratory, evidence maturity, resource allocation, governance model, decision rights, metrics, transition path, time-to-learning, time-to-impact, and risk. The agent classifies initiatives by strategic role, maturity, uncertainty, and evidence rather than by a simple date, then checks whether the organization is protecting the core, scaling the next engines, and creating enough credible future options at the same time."
  # purpose: problem this skill solves.
  purpose: "This skill prevents agents from turning innovation strategy into either short-term optimization or distant invention theater. It forces the agent to show whether the portfolio funds current performance, builds emerging growth, maintains future options, uses different metrics and governance by horizon, aligns resources and incentives with strategy and risk appetite, accounts for the speed at which disruption can now arrive, and has plausible transition or shutdown paths for initiatives to move toward material business impact."
  # boundary: what this concept is not.
  concept_boundary: "Three Horizons is for innovation and growth portfolio balance across current core, emerging businesses, and future options. It is not BCG growth-share portfolio allocation, Ansoff product-market growth path selection, Blue Ocean value-curve redesign, scenario construction and stress-testing across alternative external futures, OKR goal-setting, expected-value calculation, generic roadmap sequencing, or the International Futures Forum social-change Three Horizons method unless the user explicitly asks for futures or systems-change facilitation."
  # analogy: one-sentence metaphor preserving the mechanism.
  analogy: "Three Horizons is like managing a farm: harvest today's crop, cultivate the next field, and plant experiments for future seasons without pretending one task can wait until the others are finished."
  # misconception: common wrong mental model and correction.
  misconception: "The common mistake is treating Horizon 1, Horizon 2, and Horizon 3 as now, later, and much later work that should be handled sequentially. The framework's point is concurrent management: all three horizons need attention now, but with different evidence standards, resources, metrics, and governance. Two further errors compound it: in fast-moving markets the time axis collapses entirely and a Horizon 3 disruption can be delivered on a Horizon 1 timeline; and the S-curve from H3 to H2 to H1 is a stylized simplification, since some H3 options skip H2 and hit H1 directly while capital-intensive ones need a large step-change before any gradual curve appears."
# relations: legacy plain-SKILL.md compatibility surface; canonical graph edges live under metadata.relations.
relations:
  related: ["ansoff-matrix","bcg-matrix","blue-ocean-strategy","scenario-planning","playing-to-win","expected-value","okrs","swot-tows","value-chain-analysis","epistemic-grounding","methodology"]
  suppresses: ["bcg-matrix","ansoff-matrix","blue-ocean-strategy","scenario-planning","okrs","expected-value","playing-to-win"]
  verify_with: ["epistemic-grounding","methodology","expected-value"]
---

## Concept of the skill

**What it is:** Three Horizons is an innovation and growth portfolio framework for balancing current core performance, emerging businesses, and future options. It helps an organization manage Horizon 1, Horizon 2, and Horizon 3 work concurrently rather than sacrificing future growth to near-term pressure or treating exploratory ideas as a disconnected lab.

**Origin:** The framework was introduced by McKinsey consultants Mehrdad Baghai, Stephen Coley, and David White in *The Alchemy of Growth* (1999). Their core empirical finding was that durable high growth comes not from bold leaps but from a *staircase* of measured, overlapping steps — companies that sustained growth were always building the next engine while running the current one. Three Horizons is the portfolio expression of that staircase; it is not a sequential "do H1 now, H2 later, H3 someday" ladder.

**Mental model:** Treat growth as a portfolio of overlapping maturity curves. Horizon 1 funds and extends the current core; Horizon 2 turns promising opportunities into scaled engines; Horizon 3 creates options for future businesses, technologies, capabilities, or models that are still uncertain. The S-curve progression from H3 through H2 to H1 is a stylized simplification — real transitions are often nonlinear, with some H3 options skipping H2 and disrupting H1 directly, and others requiring large upfront capital before any gradual curve appears.

**Why it exists:** Agents often collapse innovation strategy into a roadmap, a list of ideas, a theater of distant invention, or a single investment decision. This skill forces portfolio balance, actual resource evidence, different governance by horizon, evidence fit, explicit transition paths, speed-of-disruption checks, and honesty about kill criteria and budget protection.

**What it is NOT:** It is not BCG, Ansoff, Blue Ocean Strategy, scenario planning, OKRs, expected-value math, a fixed time-phased roadmap, or the futures-studies Three Horizons facilitation method unless the user asks for that variant.

**Adjacent concepts:** innovation portfolio, corporate growth strategy, current core, adjacent growth, transformational bets, the Innovation Ambition Matrix (core / adjacent / transformational), explore/exploit portfolio, venture pipeline, R&D portfolio, option value, innovation accounting, staged funding, ring-fenced budget, governance, portfolio metrics, transition risk, time-to-impact.

**One-line analogy:** Three Horizons manages the growth garden by harvesting, cultivating, and planting at the same time.

**Common misconception:** Horizon 3 is not work to ignore until the distant future. It is uncertain work that needs small, credible, current investment and learning now — and in fast-moving markets it can become a live competitive threat in months, not years.

# Three Horizons

## Domain Context

Use Three Horizons for innovation portfolio reviews, corporate growth strategy, product and R&D portfolio planning, transformation roadmaps, venture studio reviews, new-business building, and strategy memos that need to balance today's performance with future opportunity creation. Use public, aggregate, or synthetic examples only. Do not include private strategy data, customer data, payment details, deal details, employee-level facts, secrets, or confidential financials in examples or evals.

The framework is strongest when the user asks whether a portfolio is too incremental, whether future bets are credible, whether emerging businesses are getting enough investment, how to govern different kinds of initiatives, how to stop current-core pressure from crowding out future growth, or how actual funding and leadership attention compare with a stated growth ambition. It is weaker when the question is about industry attractiveness, product-market growth quadrant selection, market-boundary reconstruction, constructing alternative future worlds, quarterly execution metrics, or a single quantified investment choice.

Do not treat the three horizons as a simple time sequence. McKinsey's own framing says companies manage all three simultaneously. The time axis describes how ventures *may* mature, not when leaders should start paying attention.

**The collapsed-time caveat (modern critique).** Horizons are frequently read as fixed delivery windows — H1 ~0-12 months, H2 ~2-3 years, H3 ~3-6 years — but that calendar reading is an interpretation many later users imposed, not a fixed rule of the original framework. McKinsey's own framing is explicit that the x-axis is *not* a prompt for when leaders should start paying attention, and Baghai (a co-author) has noted the horizons were never intended as fixed timing. Steve Blank's widely cited 2019 critique ("The Fatal Flaw of the Three Horizons Model," echoed in HBR) targets exactly the organizations that *did* harden those windows into a plan: in digital, platform, and AI-disrupted markets that time assumption is not just outdated but dangerous — "Horizon 3 ideas — disruption — can be delivered as fast as ideas for Horizon 1." A competitor recombining existing technologies (Blank's example is Uber: existing smartphones plus existing drivers in a new business model) can ship a Horizon 3 disruption on a Horizon 1 timeline when enabling technology, distribution, capital, or partner ecosystems already exist. The strategic risk is that the timeline reading lulls incumbents into treating disruption as years away when it is quarters away. When you apply this skill, classify by maturity and uncertainty — never by assumed delivery date — separately assess time-to-learning and time-to-impact, and explicitly flag any Horizon 3 threat whose enabling technology already exists and could be deployed quickly against the core.

## Coverage

This skill teaches agents to:

1. Frame the portfolio owner, strategic ambition, business boundary, risk appetite, and decision the portfolio review must inform.
2. Classify initiatives into Horizon 1 current core, Horizon 2 emerging growth, and Horizon 3 future options.
3. Separate calendar timing from maturity, uncertainty, evidence, governance, strategic role, time-to-learning, and time-to-impact.
4. Check whether current performance, emerging business building, and option creation are all being funded and managed.
5. Compare declared strategy with actual funding, talent, leadership attention, and decision rights, and surface where they contradict.
6. Match metrics, decision rhythm, talent model, funding model, and governance to horizon uncertainty rather than using one operating model for all work.
7. Apply resource-allocation benchmarks (the 70-20-10 heuristic) as a starting point calibrated to context, not as a fixed law, and account for the inverse-returns pattern.
8. Account for collapsed time: assess how fast a Horizon 3 disruption could actually reach the core, and whether the portfolio is fast enough across all horizons.
9. Layer an evidence-quality overlay on horizon labels and protect ring-fenced H2/H3 budgets from H1 reallocation pressure.
10. Diagnose starvation, gap, speed, zombie-portfolio, transition, time-to-impact, and innovation-theater risks.
11. Convert the portfolio view into rebalance actions: protect, extend, invest, incubate, meter, ring-fence, accelerate, partner, harvest, pause, kill, or reclassify.
12. Distinguish Three Horizons from BCG, Ansoff, Blue Ocean Strategy, scenario planning, OKRs, expected-value analysis, and futures facilitation.

## Philosophy of the skill

Three Horizons is useful because current businesses are loud. They have customers, revenue, managers, dashboards, and urgent problems. Future options are quiet. They are uncertain, easy to underfund, and often judged with the wrong metrics. Horizon 2 work is especially vulnerable: it is too speculative for core-business governance but too concrete to remain a research project — the "middle child" of the portfolio.

The skill therefore treats the framework as a governance and portfolio-balancing method, not a decorative three-column slide. A useful Three Horizons analysis does not merely label initiatives. It asks whether each horizon has the right ambition, evidence, resources, metrics, governance, decision rights, and transition path.

The portfolio view must follow the money and attention. If the strategy says growth depends on new platforms but nearly all funding, executive attention, and senior talent remain locked in near-term optimization, the Three Horizons answer should name that contradiction. Conversely, a portfolio that overfunds speculative ideas while neglecting core cash flow is not visionary; it is starving the engine that funds future options. Check the incentive layer too: even a balanced *budget* will starve H2/H3 if promotion, compensation, and recognition systems reward only short-term core outcomes, because talent quietly migrates back to the work that gets rewarded.

Nagji and Tuff's HBR work on innovation portfolios is useful as a starvation warning: in their 70/20/10 core/adjacent/transformational pattern, the small transformational share tended to produce a disproportionate share of longer-term value — roughly the inverse of its resource share. Treat that as directional evidence, not a quota. It explains why H3 options should not be dismissed as "only 10% work," while still preserving the rule that resource ratios are diagnostics, not universal targets.

The framework's own weakness is the time axis. Read literally, it implies disruption is far away and can be planned for at leisure. The corrected reading keeps the three strategic roles (run the core, build the next engine, create future options) but drops the assumption that horizon equals calendar distance. The discipline is to manage all three now, at horizon-appropriate evidence standards, and to move with enough speed that a fast-arriving disruption does not catch the core flat-footed.

## Workflow

### 1. Frame the portfolio question

Start by naming the portfolio and the decision this analysis must support.

```text
Portfolio owner:
Strategic ambition:
Business, product, or market boundary:
Decision this analysis must inform:
Current core performance pressure:
Known growth gap:
Risk appetite:
Disruption exposure (how fast could a competitor reach our core?):
Initiatives in scope:
Evidence available:
Actual funding / talent / executive attention / decision rights:
Target allocation split (e.g., 70/20/10 or custom):
Are H2/H3 budgets ring-fenced from H1 reallocation?:
Resource constraints:
Privacy boundary:
```

If the user only provides a list of initiatives, ask what decision they need: rebalance funding, identify gaps, set governance, review roadmap risk, test resource alignment, or choose which initiatives to advance.

If the user has no strategic ambition or where-to-play / how-to-win direction, state that Three Horizons can diagnose portfolio shape but cannot invent the strategy by itself. Use `playing-to-win` for upstream strategy-choice work.

### 2. Define the three horizons

Use the horizons as strategic roles, not just dates.

| Horizon | Strategic role | Typical work | Evidence standard | Common governance |
| --- | --- | --- | --- | --- |
| Horizon 1 | Extend, defend, renew, and improve the current core that produces most current profit or cash flow | core product improvements, operational efficiency, pricing, channel expansion, customer retention, renewal work that protects the core from decline | strong performance data, operating metrics, known customers, clear financial impact, business-model fit | business-line ownership, operating reviews, near-term financial and customer metrics |
| Horizon 2 | Build emerging opportunities that could become meaningful growth engines | adjacent businesses, new products with traction, new channels, scale-up ventures, new customer segments, business-model extensions | market traction, repeatable unit economics, capability fit, product-market fit, scaling risks | dedicated growth governance, staged funding, scale milestones, cross-functional ownership, protection from premature core metrics |
| Horizon 3 | Create options for future growth under high uncertainty | research, prototypes, minority stakes, exploratory ventures, new technologies, new capabilities, new business-model experiments | assumptions tested, learning velocity, signal quality, problem-solution fit, strategic option value, cheap invalidation | small bets, discovery governance, metered funding, venture-board oversight, option reviews, kill or continue based on learning |

Horizon 1 is not bad and Horizon 3 is not automatically visionary. The question is whether the portfolio has enough of each for the strategy, environment, risk appetite, and time-to-impact. Note that *horizon is a role, not a date*: a Horizon 3 disruption can arrive on a Horizon 1 timeline if its enabling technology already exists.

**Horizon 2 is the bridge.** It connects the current core to future options and is the most frequently underfunded and misgoverned horizon. Concrete example: Amazon Web Services spent years as an H2 bet inside Amazon before it became a Horizon 1 core. Without dedicated ownership, staged funding, and protection from core-business metrics, H2 initiatives either stall or get prematurely scaled. Within H2 it helps to separate *sustaining* emerging bets that make the current model better from *transformative* emerging bets that pave the way for a new model — both are legitimate, but they carry different risk and need different protection.

### 3. Classify initiatives by role, evidence, and uncertainty

For each initiative, record why it belongs in a horizon.

```text
Initiative:
Proposed horizon:
Strategic role:
Customer or market evidence:
Business-model evidence:
Capability or technology maturity:
Uncertainty type:
Evidence quality (strong / partial / weak / none tested):
Time-to-learning:
Time-to-impact:
Current investment:
Talent and leadership attention:
Expected value path:
Governance owner:
Metric that should decide next funding:
Transition condition:
Kill / pause condition:
```

Do not classify only by launch date. A project launching next quarter can still be Horizon 3 if the business model is unproven. A current product may be Horizon 1 even if it has a multi-year roadmap. A disruptive H3 threat can be near-term if the enabling technology and distribution already exist.

Use these tests when the classification is fuzzy:

| Test | H1 signal | H2 signal | H3 signal |
| --- | --- | --- | --- |
| Evidence maturity | business-model fit: predictable unit economics, stable model | product-market fit: customer pull, repeatable sales | problem-solution fit: assumptions tested, prototypes, early signals |
| Customer certainty | known customers and behavior | early traction from target customers | problem, customer, or use case still being discovered |
| Business model | current model works | model is plausible and being scaled | model is hypothetical or intentionally optional |
| Capability gap | current capability system can deliver | new capability must be built or integrated | capability, technology, or market logic is still uncertain |
| Funding logic | performance return and renewal | staged investment toward scale | metered learning and option preservation |
| Governance owner | core business owner | dedicated growth owner or sponsor | discovery owner with explicit learning mandate |

Layer an **evidence-quality overlay** on top of the horizon labels: an H2 project with weak evidence should be governed more like H3, and an H3 project whose validated-learning milestones exceed targets should be considered for H2 transition. Two projects in the same horizon with very different evidence quality are not the same bet.

### 4. Perform the portfolio teardown

Look at the whole portfolio, not only the labels. Follow where resources and authority actually flow.

| Teardown question | Why it matters |
| --- | --- |
| Where do dollars, talent, executive attention, and decision rights actually go? | Portfolio balance is not real unless resources follow it. |
| Is spending aligned with strategic objectives and expected growth areas? | Innovation portfolios often keep funding work after strategy has shifted. |
| Does the risk profile match risk appetite and ambition? | Conservative portfolios underproduce new growth; reckless portfolios can starve the core. |
| Which projects are frozen, zombie, or politically protected? | Stalled work consumes resources and blocks better options. |
| Who can pause, kill, unfreeze, or scale work? | Decision rights decide whether the portfolio can actually rebalance. |
| Are H2/H3 budgets ring-fenced from H1 reallocation pressure? | An H3 ambition with no protected allocation is not a real H3 portfolio; it gets raided under quarterly pressure. |
| Which horizon is measured with the wrong metric? | Core metrics can kill exploration; exploration looseness can excuse weak H1 execution. |

Name resource allocation explicitly. A portfolio that claims Horizon 3 ambition but assigns no funding, owner, or learning metric does not have a Horizon 3 portfolio.

### 5. Check portfolio balance

Look for patterns across the whole portfolio.

| Pattern | What it means | Diagnostic question |
| --- | --- | --- |
| Horizon 1 overweight | Current performance dominates future creation | Are core initiatives consuming all leadership attention, senior talent, and funding? Is actual allocation far from the target split? |
| Horizon 1 neglected | The cash engine or customer trust is being damaged while the organization chases novelty | Is current performance strong enough to fund and legitimize future work? |
| Horizon 2 gap | There are ideas and core work, but few scale-ready growth engines | Which H3 options have evidence enough to become H2, and what blocks the transition? |
| Horizon 2 orphaning | Emerging opportunities have traction but no durable owner, channel, operating model, or funding path | Who is accountable for turning traction into a scalable business? |
| Horizon 3 theater | Exploratory ideas exist but no assumptions are tested | What learning, option, or signal would justify continued investment? What is the kill rate for H3? |
| Disruption blind spot / fast H3 threat | A disruptive option is treated as distant even though it can reach the core quickly | Could a competitor recombine existing technology, distribution, or partners to hit our core within a year? |
| Same metrics across horizons | Core-business control is imposed on exploration, or exploration looseness is imposed on core | Which metric fits each horizon's uncertainty and maturity? |
| No transition logic | Initiatives sit in columns without movement | What evidence moves an initiative from H3 to H2, from H2 to H1, or to shutdown? |
| Funding mismatch | Resource allocation contradicts declared strategic ambition | Where do dollars, talent, executive attention, and decision rights actually go vs. the target split? |
| Incentive misalignment | Funding looks balanced, but promotion, compensation, and recognition reward only H1 delivery, so the best people avoid H2/H3 | Do career advancement and bonuses ever reward a failed-but-well-run H3 experiment, or only shipped H1 revenue? |
| No ring-fenced budgets | H2/H3 budgets are vulnerable to H1 reallocation | Are H2/H3 budgets protected at the board level? Do they survive quarterly earnings pressure? |
| Low kill rate or high zombie index | Governance is not making hard decisions | How many initiatives were killed last period? How many are still active despite stalled evidence? |

Name resource allocation explicitly. A portfolio that claims Horizon 3 ambition but assigns no funding, owner, or learning metric does not have a Horizon 3 portfolio.

### 6. Set the resource-allocation benchmark

Before recommending specific moves, sanity-check the *split* of resources against a known benchmark — then adjust to context.

The most cited empirical anchor is the **70-20-10 heuristic** from Bansi Nagji and Geoff Tuff's *Managing Your Innovation Portfolio* (HBR, 2012). Across the companies they studied, the high-performing pattern allocated roughly:

- **70%** of innovation resources to **core** (Horizon 1) — improving existing offerings for existing customers.
- **20%** to **adjacent** (Horizon 2) — expanding into related markets, customers, or offerings.
- **10%** to **transformational** (Horizon 3) — creating entirely new offerings or markets.

Two findings make this more than a budgeting rule:

1. **The returns invert.** Nagji and Tuff observed that the ~10% spent on transformational work tended to generate a disproportionate share (they cite roughly 70%) of long-term value creation, while core work delivered steady but bounded returns. A portfolio funded only where the near-term return is safest systematically underfunds where the long-term value is.
2. **It is a heuristic, not a law.** The right ratio depends on industry maturity, disruption exposure, cash position, ambition, and capability. A stable, cash-rich incumbent in a slow industry may sit near 70-20-10; a firm facing fast disruption may tilt toward H2/H3 (for example, nearer 50/30/20 in a high-disruption technology sector); a capital-constrained startup may run closer to 90/10/0 until the core is stable. Use the benchmark to *detect imbalance* (e.g., "you claim transformational ambition but allocate 2% to it"), not to prescribe a fixed number — and do not convert it into a hard control rule such as an automatic alert at a fixed deviation.

Always state allocation in terms the organization actually controls — money, talent, executive attention, and decision rights — not slide real estate. State what the current allocation actually is, what the strategic ambition implies it should be, and the gap plus the action to close it.

### 7. Match governance and metrics to each horizon

Use different controls for different uncertainty levels.

| Horizon | Good metrics | Bad metric fit | Funding and decision rhythm |
| --- | --- | --- | --- |
| H1 | revenue, margin, retention, productivity, quality, customer satisfaction, cash flow, renewal progress | only learning milestones, no financial accountability | operating cadence; monthly or quarterly; fund for performance, renewal, and resilience |
| H2 | traction, repeatability, unit economics, adoption, capability readiness, scale bottlenecks, sponsor commitment | mature-core profit thresholds too early, vanity pilots, endless protected pilots | staged funding; milestone reviews; scale when repeatability improves |
| H3 | assumptions tested, learning speed, signal quality, option value, strategic relevance, cheap invalidation, time-to-learning, willingness-to-pay signals, partner traction | revenue targets before the model exists, indefinite exploration with no kill criteria, demos with no assumptions tested | metered funding; small bets first; venture-board oversight; continue, pivot, pause, or kill based on evidence |

If one governance model is applied to all horizons, call out the distortion. H1 needs discipline and performance — a steering-committee posture. H2 needs scaling evidence and protection from premature core metrics. H3 needs fast learning, cheap invalidation, explicit option logic, and a venture-board posture that funds in tranches tied to validated learning rather than annual budget entitlement.

Track a few **portfolio-level governance-health metrics** alongside the per-horizon ones:

| Metric | What it measures |
| --- | --- |
| Portfolio balance ratio | Actual vs. target allocation across H1/H2/H3 |
| Kill rate | Initiatives killed per period / total active — too low means governance is not making hard calls |
| Zombie index | Stalled initiatives with no recent evidence / total active |
| Vitality index | Revenue from innovations launched in the last N years / total revenue |

### 8. Diagnose transition risks

The portfolio is useful only if initiatives can move or stop.

| Transition | Risk | What to check |
| --- | --- | --- |
| H3 to H2 | Promising option cannot find a business owner, capability path, customer evidence, or funding model | sponsor, problem evidence, customer evidence, business-model hypothesis, required capability, risk-appetite fit |
| H2 to H1 | Emerging business is scaled before repeatability or starved before scale | unit economics, operating model, channel readiness, support model, leadership ownership |
| H1 renewal | Core business optimization blocks cannibalization, category shifts, or platform renewal | incentives, customer migration, product architecture, sales conflict, margin dependency |
| H3 fast threat to H1 (skipping H2) | Disruption is classified as distant while competitors can deploy it quickly, or it attacks the core before the organization is ready | enabling-technology availability, distribution access, switching costs, competitor incentives, organizational resistance |
| Capital-intensive transition | Initiative needs a large upfront commitment before any gradual scaling is possible | tooling cost, regulatory-approval timeline, manufacturing scale threshold |
| Shutdown | Weak initiatives continue because they are politically protected | kill criteria, opportunity cost, evidence quality, owner incentives, resource-redeployment path |

Name what evidence would change the classification. Without transition criteria, the horizon labels become static decoration.

### 9. Recommend rebalance actions

The output should be a short portfolio diagnosis followed by concrete moves.

| Action | Use when |
| --- | --- |
| Protect | H1 cash flow, customer trust, operational resilience, or regulatory reliability is at risk and future work depends on it |
| Extend | H1 has overlooked near-term growth, renewal, channel, pricing, or efficiency opportunities |
| Invest | H2 has evidence and needs scale resources, dedicated ownership, or executive protection |
| Incubate | H3 has strategic option value but needs cheap learning, not scale funding |
| Meter | H3 or early H2 needs staged investment tied to evidence rather than annual budget entitlement |
| Ring-fence | H2 or H3 budget, talent, or reporting line needs protection from H1 reallocation pressure |
| Accelerate | a fast-arriving H3 threat or opportunity needs deployment speed, not study |
| Partner or acquire | capability, access, speed, or risk sharing cannot be built internally in time |
| Harvest / sunset | an H1 business is in structural decline and is no longer strategic; manage it for cash with a sunset plan and migrate customers to newer platforms |
| Diversify | the H3 portfolio has too many correlated bets; add uncorrelated options |
| Pause or kill | evidence is weak, learning has stalled, strategy has shifted, or opportunity cost is too high |
| Reclassify | the initiative's maturity, evidence, or role does not match its label |

Do not recommend an even split by default, and do not recommend a fixed 70-20-10 split either. The right balance depends on industry maturity, disruption risk, cash position, ambition, capability, time to impact, and risk appetite. Do not let rebalancing become a vague "more innovation" recommendation — state what moves: funding, senior talent, leadership attention, decision rights, governance cadence, metrics, partner path, or shutdown authority.

## Modern Adaptations

Practitioners now run Three Horizons with several adjustments to the 1999 original. Fold these in when relevant:

- **Horizons as maturity/uncertainty bands, not timeframes.** The durable insight is the three strategic roles; the calendar timings were always rough, were never the framework's intent, and are now misleading. Stage-gate each horizon on evidence and learning, not elapsed time. In fast-moving markets the traditional decades-long S-curve compresses to a few years, so all three horizons increasingly share the same decision window.
- **Speed across all horizons (Blank's prescription).** Because disruption can arrive fast, the corrective is not to abandon the model but to compress cycle time everywhere — favor "speed of good-enough deployment" for H3 bets that could otherwise be overtaken, and watch for competitors recombining existing technology against your core.
- **Combine with real-options and lean discovery (innovation accounting).** H3 work is option-creation: many small, cheap, invalidate-early bets, governed on learning velocity with explicit kill criteria and metered funding tranches tied to validated-learning milestones — not one big multi-year program. Manage H3 like a venture portfolio: diversify, expect most bets to fail, and require the portfolio to produce enough winners.
- **The Innovation Ambition Matrix as a complementary lens.** Nagji and Tuff's core/adjacent/transformational matrix maps the same three bands onto two axes (where to play: existing vs. new markets; how to win: existing vs. new products/assets). It is a useful refinement when the user wants to place an initiative more precisely, but the portfolio-balance, governance, and transition logic remain the job of this skill.

## Limitations and Caveats

Name these known weaknesses rather than concealing them; each has a mitigation the agent should apply:

| Limitation | Mitigation |
| --- | --- |
| No built-in kill criteria | Supply explicit kill criteria, track kill rate, and flag zombie projects. |
| S-curve assumption masks real transition patterns | Note that some H3 options skip H2 and that capital-intensive industries may require step-change investment with no gradual curve. |
| Time-based classification can hide evidence quality | Layer an evidence-quality overlay (strong/partial/weak/none) on top of horizon labels. |
| No industry variation in the base framework | Adapt cadences, capital assumptions, and metrics to the specific industry (software, manufacturing, life sciences). |
| Can be misused as a decorative slide without resource allocation | Verify that actual resource allocation matches the declared horizon split. |
| May create a false sense of predictability | Communicate that horizon timing is inherently uncertain; the framework is a governance tool, not a prediction engine. |

## Output Template

Use this compact structure when applying the framework.

```text
Three Horizons diagnosis:

Portfolio boundary:
Strategic ambition:
Risk appetite:
Current portfolio pattern:
Resource reality (funding / talent / executive attention / decision rights):
Target vs. actual allocation (e.g., 70/20/10):
Speed / disruption exposure:

Horizon 1 - current core:
- Initiatives:
- Evidence:
- Resource level:
- Governance / metrics:
- Risks:
- Recommended action:

Horizon 2 - emerging growth:
- Initiatives (sustaining vs transformative):
- Evidence:
- Resource level:
- Governance / metrics:
- Scale bottlenecks:
- Transition criteria:
- Recommended action:

Horizon 3 - future options:
- Initiatives:
- Evidence:
- Resource level:
- Learning metrics:
- Time-to-learning / time-to-impact:
- Kill or continue criteria:
- Recommended action:

Portfolio governance:
- Kill rate (last period):
- Zombie index:
- Ring-fenced H2/H3 budgets in place?: yes / no / partial
- Evidence-quality overlay applied?: yes / no

Cross-horizon risks:
- Starvation:
- Gap:
- Speed / disruption blind spot:
- Governance mismatch:
- Metrics mismatch:
- Incentive misalignment:
- Transition bottleneck:
- Zombie or politically protected work:

Next decisions:
1.
2.
3.
```

## Boundary With Nearby Skills

| Nearby skill | Use that skill when | Use Three Horizons when |
| --- | --- | --- |
| `bcg-matrix` | The portfolio question is market growth x relative market share across business units or products | The question is innovation/growth maturity, evidence, governance, and current-vs-future balance |
| `ansoff-matrix` | The question is which product-market growth path an option represents | The question is how options across maturity levels balance in a growth portfolio |
| `blue-ocean-strategy` | The task is value innovation, strategy canvas, ERRC grid, or market-boundary reconstruction | The task is balancing current, emerging, and future growth initiatives |
| `scenario-planning` | The task is constructing alternative external future worlds from critical uncertainties and stress-testing strategy across them | The task is balancing the internal portfolio across current core, emerging growth, and future options |
| `playing-to-win` | The team needs an integrated strategy cascade before choosing portfolio bets | The strategy direction exists and the question is portfolio balance and governance |
| `expected-value` | Outcomes, probabilities, and values are estimable enough for quantitative option comparison | The task is portfolio-level classification, evidence fit, and governance design before valuation |
| `okrs` | The team needs measurable objectives and key results for execution | The team needs to decide what kinds of innovation work should exist and how they should be governed |

## Verification

Before giving the final analysis, check:

- Did you classify by strategic role, maturity, uncertainty, and evidence rather than by date alone?
- Did you explain that all three horizons require concurrent management?
- Did you assess how fast a Horizon 3 disruption could reach the core (collapsed-time check), not assume it is years away?
- Did you include Horizon 1 value and cash flow instead of treating the core as merely obsolete?
- Did you identify whether Horizon 2 is underfunded, overprotected, orphaned, or missing?
- Did you give Horizon 3 learning metrics, metered funding logic, and kill or continue criteria rather than vague ideation?
- Did you sanity-check the resource split against a benchmark (e.g., 70-20-10) while treating it as a heuristic, not a law, and note the inverse-returns pattern?
- Did you compare actual resources with declared strategic ambition and risk appetite?
- Did you check whether incentives, promotion, and recognition reward only H1 outcomes and so starve H2/H3 even when the budget looks balanced?
- Did you assess whether H2/H3 budgets are ring-fenced or vulnerable to H1 reallocation?
- Did you layer an evidence-quality overlay (strong/partial/weak/none) on the horizon labels?
- Did you match governance and metrics to horizon uncertainty?
- Did you name transition criteria between horizons and shutdown criteria for stalled work?
- Did you avoid private customer, employee, financial, or strategy details?

## Do NOT Use When

| Use instead | When |
| --- | --- |
| `bcg-matrix` | Portfolio allocation is based on market growth and relative market share |
| `ansoff-matrix` | The task is classifying a growth move by existing/new products and markets |
| `blue-ocean-strategy` | The task is reconstructing market boundaries or designing a new value curve |
| `scenario-planning` | The task is constructing alternative future worlds with signposts, hedges, and contingencies and stress-testing strategy across them |
| `okrs` | The task is writing execution goals and key results |
| `expected-value` | The task is comparing quantified scenarios by probability and payoff |
| `swot-tows` | The task is inventorying internal/external factors and generating options |