Three Horizons Analysis

The Agency That Optimized Its Way Into Obsolescence

Read the investment allocations of a mature space agency on any particular technology domain and a pattern usually surfaces. The overwhelming share of current-year spending is committed to incremental improvement of the operational workhorse — the chemical propulsion architecture whose reliability is trusted, whose supplier base is qualified, whose regulatory pathway is clear. A smaller share is committed to the emerging alternative that has reached mid-level maturity and is beginning to scale on science missions. A vanishing share, often rounded to zero in practice, is committed to the frontier option that remains mostly conceptual and whose payoff, if any, is years distant.

Every budgetary signal justifies this allocation. The operational workhorse is producing results now. The emerging alternative has not yet demonstrated the cost parity that would justify reallocation. The frontier option’s performance is speculative. Any year’s budget defense can be made to read as prudent. And yet, across a decade of such budgets, the same agency finds itself structurally behind the technology curve its competitors have climbed. The frontier option that the agency neglected has matured into the emerging alternative that its competitors are now scaling. The competitors’ emerging alternative has become their operational workhorse. The agency’s operational workhorse is still operational — and now outpaced.

This is the “optimization trap” that three horizons analysis was designed to expose. The method’s claim is not that emerging or frontier technologies are always the right bet; its claim is that a portfolio that cannot read its own allocation across horizons cannot recognize when the allocation has become strategically self-defeating.

Baghai, Coley, and White — Then Bill Sharpe

The three horizons framework has two parents whose contributions are distinct and complementary. Its first articulation appeared in The Alchemy of Growth (1999), written by Mehrdad Baghai, Stephen Coley, and David White out of McKinsey’s strategy practice. Their concern was corporate growth pipelines, and their observation was that companies sustaining growth over multiple decades did not simply milk a single business line but maintained a deliberate balance across three concurrent horizons of maturity: core businesses defending current earnings, emerging businesses scaling toward the next generation of revenue, and options businesses exploring transformations that might define a future decade. Their formulation was implicitly sequential — horizons as staggered time periods through which a single business would travel — but its enduring contribution was the portfolio logic: the argument that an organization’s strategic health is legible not from any single initiative but from the balance across initiatives at different maturities simultaneously.

Bill Sharpe extended the framework materially in Three Horizons: The Patterning of Hope (2013), drawing from futures practice rather than corporate strategy. Sharpe’s key reformulation was to treat the horizons as coexisting patterns rather than sequential time periods. At any given moment, H1 is what is currently dominant, H2 is what is already emerging, and H3 is what is beginning to appear at the margins of the system — and they are always all three present at once. Sharpe’s formulation made the framework more usable for technology landscape analysis, because it reframed the question from “what will we do in five years” to “what is already present at each level of maturity, and how should we read the transitions between them.”

The space-sector adoption has drawn on both lineages. The McKinsey portfolio logic is visible in how agencies and companies allocate R&D investment across horizons. Sharpe’s coexistence logic is visible in how technology landscape analyses treat the field — not as a sequence of dominant regimes but as a simultaneous configuration of mature, emerging, and frontier technologies whose transitions are the strategic content. Modern practice uses both, and practitioners who hold only one parent in view typically miss half the method.

What the Horizons See That a TRL Ladder Does Not

The characteristic analytical gesture of three horizons analysis is its substitution of strategic posture for maturity rating. A TRL scale tells the practitioner where a technology is on a one-dimensional maturity axis. A horizons map tells the practitioner what to do about it.

H1 — Defend and Extend — names the technologies currently delivering value, mature, approaching the plateau of their S-curve. The strategic posture is optimization, position defense, extraction of remaining value, and disciplined management of the eventual decline. The risk profile is low on technical grounds and high on disruption grounds, because H1 is exactly where displacement by H2 or H3 actually occurs. An H1 technology optimized heavily without attention to the H2 that is approaching it is a technology being set up for a harder transition later.

H2 — Build and Scale — names the technologies transitioning from demonstration to operational deployment, growing rapidly, attracting investment, beginning to displace the H1 incumbents. The strategic posture is acceleration of scaling, resolution of integration challenges, and early market position capture. The risk profile is moderate on technical grounds and dominant on execution grounds: the question is rarely whether H2 technologies will work but whether they will scale on the timeline that investment plans assume.

H3 — Explore and Transform — names the technologies in early research, experimentation, or conceptual stages. High uncertainty, potentially transformative, often never maturing, and occasionally redefining an entire landscape when they do. The strategic posture is monitoring, exploratory bets, and the maintenance of optionality — not commitment, which would be premature, but informed awareness that preserves the organization’s ability to engage if and when H3 signals resolve.

The method’s analytical power comes from treating these horizons as coexisting rather than sequential, and from focusing the analysis on the transitions between them. H3→H2 transitions are triggered by successful demonstrations, funding commitments, regulatory enablement, or convergence with complementary technologies. H2→H1 transitions are triggered by cost parity, infrastructure readiness, demand pull, or standards adoption. H1 declines are triggered by disruption from H2 alternatives, regulatory obsolescence, resource exhaustion, or capability ceiling. A horizons map without this transition analysis is a snapshot; the method’s value sits in the transitions.

Portfolio balance is where the method earns its executive attention. Three characteristic imbalances produce three characteristic failure modes.

The balance is the strategic finding; no individual horizon tells the story.

Cislunar Propulsion Across Three Horizons

Consider the method applied to the in-space propulsion portfolio of a generic national space agency. H1 is conventional chemical bipropellant — the operational workhorse of the agency’s current missions, TRL 9, with decades of flight heritage, a mature supplier base, and performance that has approached its physical ceiling. The S-curve position is deep into maturity; incremental improvements are available but the slope is shallow, and the Tsiolkovsky equation imposes a ceiling no engineering optimization can repeal.

H2 is solar electric propulsion at 50-kilowatt class — demonstrated on science missions, currently scaling toward cargo tug applications, TRL 5–6. The S-curve position is in rapid growth; the technology has not yet approached its plateau on the relevant performance metrics. Scaling challenges exist (power-system integration, thermal management at higher power levels, thruster lifetime qualification), but none of them are principled blockers; they are engineering problems with defined paths to resolution. The H2→H1 transition depends on cost-per-kilogram parity with chemical propulsion for cargo missions, plausibly achievable by the early 2030s if power-system scaling progresses on current trajectories.

H3 is nuclear thermal propulsion — ground testing recently renewed, no modern flight heritage, TRL 3. The transformation potential is substantial: specific impulse roughly double that of chemical propulsion opens mission profiles that are currently infeasible, particularly for crewed transit to Mars and for high-delta-v cargo missions. The H3→H2 transition depends on a successful ground test campaign and on resolution of the regulatory pathway for nuclear launch, neither of which is certain; the earliest plausible transition is well into the 2030s, and the probability is moderate rather than high.

The portfolio balance read is where the method produces its strategic finding. The agency is heavily invested in H1 optimization — incremental improvements to chemical propulsion, extended qualification of existing engine variants, industrial-base support for legacy supplier networks. Its H2 commitment to solar electric propulsion is present but thin relative to the scaling timeline it would require to achieve H1-level deployment by the mid-2030s. Its H3 commitment to nuclear thermal is nominal — monitoring rather than exploratory investment. A competitor whose H2 commitment is stronger than the agency’s is on a path to operational 50-kilowatt-class SEP several years before the agency; a competitor whose H3 exploration is substantive is on a path to nuclear thermal capability that would redefine mission architecture by the late 2030s.

The non-obvious insight is that the agency’s current allocation looks prudent on every individual budget line — each H1 item justifies its incremental return — and yet the aggregate allocation is the optimization trap. No single line is wrong; the portfolio is wrong. Three horizons analysis is the method that makes the portfolio visible as a portfolio, rather than as a collection of individually defensible line items.

The recommended reallocation follows from the analysis. H1 optimization can absorb modest reduction without near-term consequence, because incremental gains have diminishing returns at the S-curve plateau. The freed capacity should flow to H2 scaling — specifically to the power-system integration bottleneck that is currently the binding constraint on SEP maturity. A modest H3 commitment to nuclear thermal monitoring and targeted demonstration contributes to optionality without requiring the large-ticket commitment that would be premature at TRL 3. No single reallocation is dramatic; the portfolio reshaping, in aggregate, is the strategic action.

Where It Earns Its Keep and Where It Falls Short

The method’s strength is the portfolio view. No other tool in the library forces executives to confront the aggregate allocation of attention and investment across horizons, and the exercise consistently surfaces imbalances that individual program reviews cannot see because they are each defending their own line item. For strategic planning, technology portfolio management, and R&D budget allocation, the method is the discipline that keeps organizations out of the optimization trap.

Its weaknesses are consistent with its simplification. The three horizons are a tripartite scheme imposed on a continuous maturity spectrum. Some technologies straddle boundaries, and reasonable analysts can disagree on whether a specific technology is late H3 or early H2. The method’s discipline is to make the evidence basis of each classification explicit and auditable, not to pretend to a crispness the underlying reality does not support. It favors incremental transition narratives and can underweight discontinuous disruption — the technology that leaps from H3 to H1 without a prolonged H2 phase, bypassing the transition logic the framework assumes. Horizon scanning is the complement that catches the black-swan signals the framework would otherwise suppress.

The method is structural rather than predictive. It reveals where technologies are, not where they will be. Temporal projections of when H3→H2 or H2→H1 transitions will occur are scenario-dependent, not forecasts; pairing with scenario planning is how the transition timing is stress-tested against divergent futures. Portfolio balance assessment assumes a single entity’s perspective and requires separate mapping for each actor in multi-actor landscapes. An analysis that conflates an agency’s portfolio with its prime contractors’ portfolios, or with the sector’s as a whole, misreads the strategic picture.

Three horizons does not replace the domain-specific methods whose depth it lacks. It supplies the organizing structure; TRL assessment supplies the maturity readings; S-curve lifecycle analysis supplies the trajectory interpretation; technology roadmapping supplies the sequencing plan; technical benchmark comparison supplies the fitness-for-purpose evaluation. A three horizons analysis used without these enrichments produces a structure without content. A three horizons analysis used with them produces the portfolio-level strategic view that the individual methods cannot supply on their own.

The library treats the framework accordingly. Horizon classifications provide the strategic framing within which roadmap milestones are sequenced. H2 and H3 technologies can be analyzed through disruption-theory lenses to assess whether the incumbents face sustaining or disruptive threats. Different scenarios imply different transition timelines, shifting which technologies reach H1 status by a target year. Trend analysis and S-curve data supply the empirical evidence for horizon placement.

For the Practitioner