Strategy Design

Three Horizons of Growth

The three horizons of growth balance the current business, emerging opportunities, and bets that could replace both.

The questionWhere will growth come from after the current business matures, and is anyone funding it?

Also called the McKinsey three horizons model, to tell it apart from a futures method of the same name.

Model · Baghai, Coley, and White, 1999

Horizon 1Core businesses. Extend and defend
Horizon 2Emerging businesses. Heavy investment first
Horizon 3Options for later growth. Keep bets small

Schematic of the three horizons of growth, drawn as three overlapping curves of profit over time.

When to use it

Quarterly results look fine, the pipeline behind the core business is thin, and nobody can say what replaces it.
A client has many innovation projects and cannot say which extend the core, which are new businesses, and which are experiments.
Budget arguments between the core and new ventures are being won by whoever has the most current revenue.
Leadership needs a shared reason for why different initiatives get different metrics, timelines, and tolerance for loss.

When to avoid it

The question is how a market might unfold under uncertainty. Scenario planning fits that, and the futures method called Three Horizons is another option.
The client is a one-product startup. There is no Horizon 1 to balance against, and a lean startup approach fits better.
Initiatives are already classified and the argument is about execution. Use a delivery method and not a portfolio model.
Someone wants a precise budget split from it. The usual 70/20/10 ratio is a rule of thumb with no empirical backing in the book.

Overview

The three horizons model sorts a company’s growth initiatives by how far each sits from today’s business. Horizon 1 is the core, Horizon 2 is emerging businesses, and Horizon 3 is options on what might come later. It supports a portfolio choice: how much attention and money each gets, and whether the core will be replaced in time.

The horizons are not a schedule. A Horizon 3 option is not something you start in five years, and all three are managed now. The time axis shows how long a business takes to matter, and how ventures move from 3 to 2 to 1. A different futures method also uses the name, and this note covers the growth version.

The horizons

Horizon 1

The core businesses, which provide most current profit and cash. The work is to extend and defend them and raise performance. The risk is that urgent core problems absorb all management attention.

Horizon 2

Emerging businesses and ventures that should produce substantial profit later and usually need heavy investment first. Judge them on growth and trajectory, not current profit. The risk is that they are held to the core's budget rules and cut for early losses.

Horizon 3

Options for later growth, such as research projects, pilots, and minority stakes. Judge them on what they teach and whether one earns a move to Horizon 2. Most will not, so keep each bet small.

What it is not

A timeline in which you work on Horizon 1 now, Horizon 2 next, and Horizon 3 later. All three are managed at once, and time describes how long a business takes to matter.
The futures method called Three Horizons, which Curry and Hodgson adapted from this model. It starts from a declining system and a preferred future, and it is a workshop method.
A required 70/20/10 allocation. The ratio is often attributed to Google, not to the model's authors, and it is one company's rule for its own portfolio.

The steps

List every initiative with a budget or an owner

Include small pilots and side projects. Ask for the investment to date and the next ask. You want everything that consumes money or attention, not only the projects leadership talks about.
Sort by distance from today's business

Place each one by how far it is from current customers, products, and capabilities. Do not sort by size or by when it pays back. Arguments over placement are useful, because they show where people disagree about what the core is.
Read the shape

Count and fund each horizon. Look for an empty or starved Horizon 2, and estimate when the core's growth flattens. The output is a picture of where growth is expected to come from, and the gap if the core stops growing.
Set rules for each horizon

Give each horizon its own metrics, governance, and tolerance for loss. Horizon 1 can use profit and return. Horizon 3 needs learning goals and small, stage-limited funding. A 70/20/10 split can serve as a starting guess to argue with, and not as a benchmark.
Review movement each cycle

Decide which options earn promotion, which ventures should merge into the core, and which should stop. Without exits, Horizon 3 turns into a permanent pile.

In practice

The chief strategy officer at a regional insurer brought thirty-one initiatives and asked which ones mattered. The team sorted them by how far each sat from current products, customers, and capabilities: twenty-two in Horizon 1, three in Horizon 2, and six in Horizon 3. Two of the six turned out to be product tweaks for existing customers and moved to Horizon 1, which left twenty-four, three, and four.

The three Horizon 2 businesses were the finding. One had run for three years on a research budget with no revenue target, and none had a named owner outside the innovation team. The core's planning cycle had no slot for them, so each fought for funding every year against claims reform. The discussion changed from which projects to add to who owns the next growth engine, and how it is funded.

What you leave with

A classified list of initiatives by horizon, with owners and funding.
A view of where growth is expected to come from over the next several years, and the gap if the core flattens.
Separate success measures and governance for each horizon.
An allocation decision with its reasoning, and not a borrowed ratio.
A list of promotion and stop decisions for the next review.

Failure points

Classification by preference

Whatever leaders like gets called Horizon 2, and unglamorous core improvements are not counted at all. The portfolio looks balanced because the labels were chosen to make it so.

Horizon 2 under core budget rules

The annual budget judges every unit on current profit. A new venture that is losing money by design loses to a claims reform program or a systems upgrade, and is cut before it would have mattered.

Horizon 3 as a pile of pet projects

Options with no stage gates and no stop rules never die. After a few years the pile holds a dozen items, and none has been tested.

One return metric for all three

A hurdle rate that suits the core removes Horizon 2 and Horizon 3 on arrival, because their returns come later and are uncertain.

Ratio as target

The 70/20/10 split gets presented as industry practice and adopted without checking what the client's core growth looks like. The numbers come from one company's rule, and they say nothing about this client.

The failures above are usually Framing: the wrong problem was chosen before this framework was applied.

Origin

Baghai, Coley, and White worked in McKinsey's growth practice. Their book The Alchemy of Growth, published by Basic Books in 1999, argued that companies which sustain growth manage three horizons at once: extending and defending core businesses, building new ones, and seeding options for the future. McKinsey republished the idea in a 2009 Enduring Ideas piece featuring Coley, which notes that the time axis shows how businesses move between horizons and is not a schedule for when to pay attention.

A different method shares the name. Andrew Curry and Anthony Hodgson adapted the model for futures work, described in a 2008 Journal of Futures Studies paper. In that version Horizon 1 is the dominant system that is losing fit with its environment, and Horizon 3 is the emerging alternative. This note covers the McKinsey growth version.