What Is the Three Horizons Model

What does the three horizons of growth model mean?

The three horizons model, also called the three horizons of growth, is a portfolio framework that sorts strategic initiatives by how far they sit from today's business. It was first set out by McKinsey's Mehrdad Baghai, Stephen Coley and David White in The Alchemy of Growth, published in 1999 according to the Internet Archive catalog record. Steve Blank's summary of the model is the clearest short version: "Horizon 1 ideas provide continuous innovation to a company's existing business model and core capabilities. Horizon 2 ideas extend a company's existing business/model and core capabilities to new customers, markets or targets. Horizon 3 is the creation of new capabilities to take advantage of or respond to disruptive opportunities or to counter disruption."

The framework exists to answer one question leadership teams struggle with constantly: how much time and budget should go toward defending today's business versus building tomorrow's. Without that explicit split, a company's own short-term incentives quietly push almost every resource toward Horizon 1, since it is the easiest work to justify in any single quarter.

Is Horizon 3 always long-term work?

Horizon 3 is defined by distance from the core, not by delivery date. In The Fatal Flaw of the Three Horizons Model, Steve Blank notes that some organizations used to define Horizon 1 as features delivered in 3 to 12 months, Horizon 2 as extensions 24 to 36 months out, and Horizon 3 as new disruptive products 36 to 72 months out. His argument is that those time bands no longer hold: "The three horizons are not bound by time." Those month ranges describe how some companies used the model, not a rule from the original book, and they still show why a time-only reading misleads: a product team can ship a Horizon 3 experiment in a few weeks if it reuses what it already has.

Why the three horizons of growth model matters for product managers

Builders Camp's Product Strategy and Head of Product certification quizzes both ask what designing strategy in horizons is for. A specific split such as 70/20/10 matters because it turns an abstract idea, balance the present against the future, into a number a leadership team can actually check its own roadmap against. A team that has quietly let 95 percent of its work fall into Horizon 1 can point to that ratio as evidence it needs to fund something further out.

Where does the 70/20/10 split come from?

The 70/20/10 ratio attached to the three horizons is usually credited to Google, not to the McKinsey authors. Blank's summary says only that "McKinsey suggested that to remain competitive in the long run a company allocate its research and development dollars and resources across all three horizons", with no percentages. Eric Schmidt's book How Google Works, as quoted on Goodreads, puts the numbers this way: "70/20/10 became our rule for resource allocation: 70 percent of resources dedicated to the core business, 20 percent on emerging, and 10 percent on new." A 2005 Business 2.0 interview with John Battelle, reproduced on the Google Operating System blog, has Schmidt describing the management version: "We spend 70 percent of our time on core search and ads."

So the honest attribution is two separate ideas that product teams now use together: the horizons from The Alchemy of Growth, and the 70/20/10 split from Google. The split was one company's rule for its own portfolio, which is a reason to treat it as a default to argue with rather than a benchmark.

What counts as Horizon 1, 2 and 3 work for a product team?

For a product team, the horizon of a piece of work depends on how far it moves from today's product and today's customers. The table below translates Blank's definitions into roadmap terms.

Horizon	What it means for a product team	Typical examples	How success is judged
Horizon 1	Improve the product current customers already pay for	Fixing the top churn reason, onboarding friction, performance, pricing page tests	Movement in retention, conversion or the North Star metric this quarter
Horizon 2	Take existing capabilities to a new segment, market or use case	A self-serve tier for smaller customers, a new vertical, a second platform	Evidence that a new segment adopts and pays, before scale
Horizon 3	Build a capability the product does not have yet	An AI agent that does the job end to end, a new business model, a platform play	Learning milestones: a validated problem, a working prototype, first usage

The practical test is the question each horizon asks. Horizon 1 asks "how do we make this better?", Horizon 2 asks "who else could this serve?", and Horizon 3 asks "what would make this product obsolete, and should we build it first?"

Is 70/20/10 the right split for a product team?

70/20/10 is a sensible starting allocation for a product with proven demand, and a poor one for a product that has not found it yet. A pre-product-market-fit startup is effectively all Horizon 1 and 3 at once, because its core is still an experiment. A mature product under pressure from a new entrant may need to push Horizon 3 well above 10 percent for a few quarters. Pick the split deliberately, write it down, and check actual capacity against it every quarter; a team that never measures its real split ends up close to 100 percent Horizon 1 by default.

How does MoSCoW work inside each horizon?

The horizons decide how much capacity each kind of work gets; MoSCoW decides what gets that capacity within each bucket. Run MoSCoW separately per horizon rather than across the whole backlog, because a single ranked list lets Horizon 1 items win every time: they have clearer value and lower risk.

Inside Horizon 1, must-haves are usually reliability and retention fixes. Inside Horizon 2, the must-have is the smallest slice that proves a new segment will adopt. Inside Horizon 3, must-haves are learning goals, not features: the one experiment that would tell you whether the bet is worth a bigger share next quarter. Won't-haves matter most in Horizon 3, where scope creep turns a cheap test into an unfunded Horizon 2 project.

How do you plan discovery, build and measurement work across horizons?

Every horizon needs all three kinds of work, in different proportions. A triple-track roadmap runs discovery, build and measurement in parallel, and laying it across the horizons shows the mix:

• Horizon 1 is mostly build and measurement, with light discovery to confirm which fix matters most.
• Horizon 2 is balanced: discovery to understand the new segment, build for the smallest adoption slice, measurement of whether that segment activates.
• Horizon 3 is mostly discovery and measurement, with build limited to prototypes and experiments until a signal justifies more.

A roadmap that shows only build work hides this. Horizon 3 looks empty on a delivery roadmap even when a team is doing exactly the right amount of discovery there.

What does a three horizons plan look like for one product team?

Take a hypothetical team of eight engineers working on invoicing software for small accounting firms, with about 100 engineer-weeks of capacity in a quarter. A 70/20/10 starting split gives roughly 70 weeks to Horizon 1, 20 to Horizon 2 and 10 to Horizon 3.

• Horizon 1 (about 70 weeks). Must: fix bank-feed sync failures, the top reason firms cancel. Should: faster month-end reconciliation. Could: a refreshed dashboard. Mostly build and measurement, tracked against monthly churn.
• Horizon 2 (about 20 weeks). Must: a trial with bookkeepers who serve freelancers, a new segment using the same invoicing core. Discovery first, then the smallest onboarding path that lets them send a first invoice.
• Horizon 3 (about 10 weeks). Must: learn whether firms would trust an assistant that drafts client invoices from timesheets. Discovery interviews plus one prototype, with a decision at quarter end on whether it earns a bigger share.

The allocation is visible, each horizon has its own must-have, and Horizon 3 has a decision date instead of an open-ended budget.

Three horizons of growth example

Builders Camp's Product Strategy practical challenge uses a B2B SaaS company with $4.2M in ARR and a Series B board decision 14 days away. The CEO wants to move upmarket into enterprise, the CTO wants to fix the core product first, and the Head of Sales wants to expand into a new vertical, marketing agencies. The challenge asks for one clear direction, a North Star metric and a board slide, not an allocation. Reading the three positions through the horizons is still useful: fixing the foundation is Horizon 1, while both the enterprise move and the agency vertical take existing capabilities to new customers, which puts both in Horizon 2 by Blank's definition. None of the three positions funds a Horizon 3 bet, and that gap is worth naming in any memo that answers the challenge.

How Builders Camp teaches the three horizons of growth model

Product Strategy, a 2 week bootcamp with 4 live sessions and 8 microlessons taught by Andre Albuquerque, covers strategic choices and trade-offs, and turning bets into a roadmap with sequencing, milestones and measurable outcomes, and uses the Series B scenario in its practical challenge. Head of Product extends the same portfolio thinking into how a leader communicates trade-offs to a board. For how the horizons connect to vision and strategy upstream, see product vision vs strategy vs roadmap; for filling the Horizon 2 and 3 buckets with differentiated options, see the Four Actions Framework. See the Product Strategy bootcamp for the full curriculum.