Three Horizons: Balancing the Core, the Next, and the Bet

Balancing the three horizons means allocating resources by risk class, not by seniority or urgency: fund Horizon 1 to defend today's revenue, Horizon 2 to grow what's next, and Horizon 3 to buy real options on the future — governed, measured, and protected as three concurrent portfolios, not three sequential phases you graduate through.

Quick answer: Run Horizon 1, 2, and 3 at the same time, each with its own metrics and governance, not as stages of maturity. Starve Horizon 3 and growth stalls three to five years out; over-fund it and today's business can't pay the bills to get there.

Three Horizons Is a Resource-Allocation Model, Not a Maturity Ladder

The Three Horizons framework, introduced by McKinsey consultants Mehrdad Baghai, Stephen Coley, and David White in their 1999 book The Alchemy of Growth, sorts a company's initiatives into three risk classes — Horizon 1 (core), Horizon 2 (next), and Horizon 3 (bet) — meant to run concurrently, funded from one shared portfolio, not as stages a business graduates through.

The framework gets misread constantly, and almost always the same way: as a funnel. Horizon 1 is "where we are," Horizon 2 is "where we're going," Horizon 3 is "someday, if there's time." That reading turns a portfolio-management tool into a roadmap slide, and it's precisely the misapplication McKinsey's own later commentary on the model has flagged — the horizons were never meant to be sequential phases a company passes through once and moves on from.

The correct read is closer to a three-fund investment portfolio than a career ladder. A pension fund doesn't wait until its bonds mature before buying equities; it holds asset classes with different risk-and-return profiles at the same time, each sized to the fund's tolerance for volatility. Horizon 3 that only gets attention "once Horizon 1 is under control" never gets attention, because Horizon 1 is never under control — it's a permanent, recurring obligation, not a problem you solve once.

This distinction matters most at the point where strategy meets funding decisions. For a deeper treatment of how advanced strategic tools like this one fit together into a coherent operating system, see this guide to advanced product strategy frameworks.

Why Horizon Three Starves First — and Over-Funding It Kills Faster

Horizon 3 starves first because it has no constituency defending it when budgets get cut — no customer calls in angry over a stalled bet, but a Horizon 1 revenue miss triggers an executive escalation within a quarter. Over-fund Horizon 3, though, and a company can die just as fast, starving the core business that pays for everything else.

That asymmetry is the whole problem: the visible failure (missing this quarter's number) always beats the invisible one (a bet quietly starving) in the room where budgets actually get decided.

Clayton Christensen's research in The Innovator's Dilemma explains the mechanism underneath the starvation pattern. Resource-allocation processes inside well-run companies are rational at every individual step — they reward proposals with clear customers, provable ROI, and short payback windows — and that rationality systematically starves anything that looks like Horizon 3, because a genuine bet cannot yet produce any of those three things.

The process isn't broken. It's optimized for the wrong horizon.

The symptoms of a starved Horizon 3 are recognizable well before revenue proves it:

• Innovation initiatives get restaffed onto Horizon 1 fire drills every time a quarterly number wobbles.
• Bet-stage projects are evaluated with the same ROI model as core features, so they always lose the comparison.
• The same handful of ambitious people rotate through every "innovation" effort, exhausted, because there's no protected headcount.
• Horizon 3 reviews get rescheduled repeatedly and eventually disappear from the calendar entirely.

Over-funding runs the opposite failure at the opposite speed. A company that pours disproportionate capital into transformational bets while its core erodes — market share slipping, retention softening, support debt piling up — doesn't get a slow decline. It gets a cash crisis, often inside twelve to eighteen months, because Horizon 1 is what generates the cash Horizon 3 is spending. Both failure modes are allocation failures, not ambition failures.

A Concrete Allocation Ratio — and When to Deviate From It

A widely used starting ratio is roughly 70% Horizon 1, 20% Horizon 2, 10% Horizon 3, drawn from Bansi Nagji and Geoff Tuff's HBR research on innovation portfolios and popularized separately by Google's engineering-time allocation rule. Treat it as a default worth stress-testing against your own industry's clockspeed and cash position, not a number to import unexamined.

Nagji and Tuff's 2012 Harvard Business Review article, "Managing Your Innovation Portfolio," analyzed corporate innovation spending across industries. It found that companies clustering near a 70/20/10 split between core, adjacent, and transformational investment tended to trade at meaningfully higher valuation multiples — commonly cited in the ten-to-twenty-percent range on price-to-earnings — than peers whose portfolios skewed heavily toward the core alone.

Google's version of the same ratio, described by Eric Schmidt and Jonathan Rosenberg in How Google Works, allocated engineer time rather than investment dollars but landed on an identical split.

McKinsey's original framework deliberately avoided prescribing a fixed ratio at all — the intended discipline was rebalancing attention and capital annually based on industry maturity and competitive pressure, not hitting a universal number.

A worked example. A 200-person B2B SaaS company at $40M ARR might land on 65% Horizon 1 (retention, reliability, expansion of the existing product), 25% Horizon 2 (an adjacent module validated with a subset of the existing customer base), and 10% Horizon 3 (two small, cheap bets outside the current category, each with an explicit kill date).

The exact split matters less than the fact that it was chosen on purpose and written down, so it can be defended against the first executive who wants to raid the smallest number for a Horizon 1 emergency.

Different Metrics, Governance, and Patience for Each Horizon

Each horizon needs its own scoreboard: judging a Horizon 3 bet on the metric that governs Horizon 1 — quarterly revenue — guarantees it gets killed before it can prove anything. Horizon 1 runs on efficiency and retention, Horizon 2 on adoption and pipeline, and Horizon 3 on learning velocity.

Those three metrics get reviewed on three different clocks — monthly, quarterly, and twice a year — with patience for Horizon 3 measured in years rather than quarters.

Rita McGrath and Ian MacMillan's discovery-driven planning approach, developed for exactly this problem, replaces a Horizon 3 project's financial forecast with a checklist of assumptions that must each be validated before the next funding tranche releases — the milestone is "we learned whether X is true," not "we hit $Y in revenue." Eric Ries's concept of innovation accounting, from The Lean Startup, applies the same logic: track validated-learning velocity as the leading indicator, because lagging financial indicators don't exist yet and forcing them into existence just produces fiction.

Horizon 1's metrics are usually the easiest to get right because they're extensions of the same friction-reduction work already tracked across the existing customer base — this complete guide to mapping the customer journey covers the emotion-and-effort curve most Horizon 1 metrics actually trace. Horizon 2's metrics are harder, because "adjacent" only means something if it's adjacent to a real, underserved job — the jobs-to-be-done framework is the sharpest tool available for confirming an adjacent market is a job worth solving before committing Horizon 2 dollars to it.

Protecting Horizon Three's Funding From Horizon One Pressure

Horizon 3 funding survives only when it's structurally insulated from Horizon 1's crisis rhythm — a ring-fenced budget line approved annually, reviewed by a separate governance forum, released against learning milestones instead of revenue, and owned by a sponsor whose performance isn't scored on the current quarter.

Without at least the first two of these, Horizon 3 funding gets reallocated the first time Horizon 1 has a bad month — which is to say, eventually, always.

Geoffrey Moore's Zone to Win offers the clearest operational model for this separation. Moore splits a company's initiatives into four zones — Performance, Productivity, Incubation, and Transformation — with Incubation deliberately excluded from the metrics and review cadence that govern the Performance Zone, so a bet's underwhelming quarter never shows up on the same dashboard as a core-product miss. That structural firewall, more than any individual person's willpower, is what keeps Horizon 3 alive through a bad Horizon 1 quarter.

Six concrete tactics that hold Horizon 3 funding in place:

1. Ring-fence the budget annually, not quarterly — an amount the CFO commits to for the fiscal year, immune to mid-year reallocation requests.
2. Give Horizon 3 its own governance forum, distinct from the core roadmap review, so it's judged against its own peers rather than against features shipping next sprint.
3. Name a single accountable sponsor, ideally a senior leader without a Horizon 1 revenue number to protect, rather than a consensus committee dominated by core-business owners.
4. Fund in tranches against assumption tests, not against a business case — each tranche is a bet on learning, not on a forecast nobody actually believes.
5. Set kill criteria before the money moves, so ending a bet is a pre-agreed outcome rather than a political fight each time.
6. Anchor the bet portfolio to a vision that outlives any single champion — a written product vision people actually repeat gives a Horizon 3 initiative a reason to survive a reorg or a leadership change, when the person who originally sponsored it is no longer in the room to defend it.

Running the Three Horizons as a Concurrent Portfolio, Not a Relay Race

McKinsey's critique of how its own framework gets applied is specific: companies treat the three horizons as a relay race, finishing Horizon 1 before handing off to Horizon 2 and then Horizon 3, when the model was designed for all three to run at once, permanently, with initiatives migrating between lanes as they mature.

A healthy portfolio always has live initiatives in all three horizons simultaneously — never zero in any one of them.

Migration between horizons is the normal life cycle, not an exception. A Horizon 3 bet that validates its core assumptions gets promoted to Horizon 2 funding and governance; a Horizon 2 initiative that proves its unit economics gets promoted into the Horizon 1 operating budget; a Horizon 1 product that's decaying gets sunset, freeing capacity for the next bet coming up the pipeline. Classification should track evolutionary stage, not calendar time or org-chart seniority.

Determining which stage a given capability is actually at is where most teams guess instead of assess. Wardley Mapping gives that judgment a visual anchor — plotting components on a genesis-to-commodity evolution axis makes it far more obvious whether a capability is still genuinely Horizon 3 territory or has quietly become table-stakes infrastructure that belongs in Horizon 1; this guide to Wardley Mapping and seeing the competitive board walks through the axis in detail.

None of this survives contact with reality if the allocation ratio lives only on a strategy slide. The gap between a tidy 70/20/10 diagram and what engineers actually build next sprint is exactly the failure mode covered in the gap between a strategy deck and daily execution — an allocation ratio has to show up in how the backlog gets scored, or it isn't really an allocation, it's a slogan.

Where Prodinja fits into horizon-based prioritization

Comparing a Horizon 3 bet against a Horizon 1 feature on the same backlog, with the same scoring model, is exactly how Horizon 3 loses every time — a near-term fix always scores higher on effort and confidence than a speculative bet, regardless of long-term value.

That's the honest use of a prioritization score: not to rank everything against everything, but to keep near-term wins from silently crowding out the long-horizon investments a portfolio actually needs.

Key Takeaways

• Three Horizons is a portfolio model, not a maturity ladder — Horizon 1, 2, and 3 initiatives should run concurrently, funded from one shared pool, not as sequential phases a company graduates through.
• Starving Horizon 3 kills slowly, typically inside three to seven years, because its lack of a revenue constituency makes it the easiest line item to raid during any Horizon 1 crisis.
• Over-funding Horizon 3 kills fast, often inside twelve to eighteen months, because Horizon 1 is what generates the cash that funds the bets in the first place.
• A 70/20/10 split (core/next/bet) is a reasonable, evidence-backed starting ratio from Nagji and Tuff's HBR research and Google's engineering-time rule — adjust it to your industry's clockspeed rather than importing it unexamined.
• Each horizon needs its own metrics, governance cadence, and patience window — judging a Horizon 3 bet by Horizon 1's quarterly revenue metric is a structural reason it dies before it can prove anything.
• Protect Horizon 3 with structure, not willpower — an annually ring-fenced budget, a separate governance forum, and pre-agreed kill criteria matter more than any individual sponsor's conviction.
• Classify by evolutionary stage, not calendar time — an initiative migrates from Horizon 3 to Horizon 2 to Horizon 1 as its assumptions get validated, and the portfolio should always have live bets in all three lanes at once.