Initiatives: why and how to use the three-horizons framework well

The three-horizons framework sorts a business's initiatives by what they are for. Horizon one extends and defends the businesses that make the money today. Horizon two builds the businesses that will be the core next. Horizon three creates the options from which the one after that will come. The framework's value is not the three labels. It is that each horizon needs different money, different measures and different people, and that a portfolio missing one of them fails in a predictable way. It was published by McKinsey consultants in 1999. In our experience it has outlived most of the tools of its era, provided it is used for the sorting and not for the timing.

This page is about using it well: why a business should sort its initiatives at all, what the sort reveals, and the ways it goes wrong.

What the three horizons are

Horizon one is the core: the businesses and products that generate most of today's profit and cash. They may still grow, but they will flatten in time, and their job in the portfolio is to fund everything else while being defended and squeezed for what they can still give. Extending and defending the core is a strategic theme in its own right, not the boring bit.

Horizon two is the emerging businesses: real activities with real revenue, growing fast, expected to become as important as today's core in the medium term. They are past the idea stage and need serious investment to get to scale.

Horizon three is the options: seeds, pilots, experiments and relationships that might become a horizon-two business. They are more than ideas, because money and people are on them, but most of them will not work, and that is expected.

Example: A food manufacturer lists its initiatives. Sorted, horizon one is the plant consolidation, the price review and the two range extensions in its main category. Horizon two is the private-label contract it won last year and is now scaling. Horizon three is a trial in a new channel and a small stake in an ingredients start-up. Nobody had noticed until the sort that horizon two was one item.

Why sort initiatives at all

Most initiative lists are flat: twenty or thirty items, unweighted, several contradicting, none clearly strategic, all competing for the same people. Sorting them by horizon forces two questions that a flat list never asks. What is this initiative for? And when, and how, is it expected to pay?

The sort also makes visible a fact that the annual budget hides. Horizon one funds the rest. If the core is not generating the cash, the other two horizons will be starved regardless of what the strategy says. The same happens when the core is asked to grow at rates it cannot deliver. Earning the right to grow comes before growing.

The deeper reason to bother is that cores do not last. Of the hundred largest listed companies in Australia in 1990, twenty-nine were still in the hundred in 2015. Most of the rest disappeared in industry consolidation, at home or abroad. A business whose whole portfolio is today's core is betting that its industry will be the exception.

And the sort exposes the unhealthy shapes. A business that is all horizon one is running out of steam and will find out in three years. A business with a strong core and a rich horizon three but nothing in between has ideas and no businesses. A business that has poured its cash into horizons two and three while the core decays has lost the right to grow. A business with none of them working is under siege. Each of these is visible in an afternoon once the initiatives are sorted, and invisible while they sit in a list.

Example: A distributor's initiative list has eighteen items. Sorted, all eighteen are horizon one: systems, cost, range, a depot. Revenue has been flat for three years and the plan shows a turn-up in year two. The momentum case and the empty horizons say the same thing.

How to use it well

Sort by purpose, not by date. The framework was originally taught with years attached to each horizon, and this is the part that has aged. In fast-moving industries a horizon-three idea can be in the market in a year; in slow ones a horizon-two business can take a decade. A public critique in 2019 argued that the time-based reading no longer applies. The framework's own authors replied that the horizons were always meant to be about role, with timelines that vary by industry. Both are right. Sort by what the initiative is for, and set its timetable from the initiative, not from the label.

Measure each horizon differently. Horizon one is judged on profit, cash, return on capital and productivity, this year. Judged on those, horizon two always loses, so it is judged instead on revenue growth, share, customer acquisition and the value it is expected to have at maturity. Horizon three is judged on milestones met, options kept open, and the rate at which ideas convert into real businesses. A single scorecard applied to all three kills the second and third horizons within two budget cycles.

Staff each horizon differently. The core needs operators who deliver the plan every month. Horizon two needs business builders who are comfortable with ambiguity and focused on the top line. Horizon three needs people who can see a market that does not yet exist. Asking the operators to build the new business is the most common staffing error, and it fails because they are good at their job.

Fund horizon two deliberately. The budget process naturally favours the core, because the core has the numbers and the horizon-two businesses have forecasts. Left to the process, horizon two ends up with token funding after business-as-usual has taken its share. The fix is to decide the split at the top, before the budget round, and protect it. In our experience a business with a healthy balance puts around half of its investment capital and senior executive time into horizons two and three. The share runs from a third to two-thirds depending on how fast the core is maturing. Most businesses are nowhere near a third.

Build staircases, not leaps. Within each theme, lay the initiatives out as a sequence of steps, each of which builds a capability the next one needs. Step one is defined in detail and ready to sign off. Steps two and three are a roadmap of decisions, not a plan, because the view changes after each step is taken and a plan written today for step three will be wrong. Staircases reduce the need for big bets and make the portfolio testable as it goes.

Check each horizon's health with plain questions. For the core: is it generating the cash to fund growth, is its cost position competitive, has its share held? For horizon two: is any new business capable of becoming as valuable as the core? Is it gaining momentum, and is the business willing to invest hard to accelerate it? For horizon three: does the leadership spend real time on it? Are the ideas on this year's list different from three years ago, and have any been turned into concrete first steps?

Kill, double or add as evidence arrives. The portfolio is not a document. Reviewed quarterly, it should change. A horizon-three trial that has met its milestones is promoted and funded, and one that has not is closed. A horizon-two business that has stalled is fixed or sold. A portfolio that has not changed in a year is not being managed.

The mistakes

Using the labels to make a list look strategic. Sorting eighteen initiatives into three columns does not make them a portfolio. If nothing is cut, weighted or owned after the sort, the framework has been used as decoration.

Treating horizon three as a parking lot. Pet projects, the founder's hobby and the initiatives nobody had the heart to kill do not belong there. Horizon three is a set of deliberate options with money, milestones and a date at which each is judged.

Expecting horizon two to fund itself. Emerging businesses need cash from the core for years before they return it. A horizon-two business asked to be profitable in year one will be run like a horizon-one business and will stay small.

Giving each function its own horizon. Marketing does not own horizon three and operations horizon one. Every leader manages across all three, and a function that has nothing beyond the core is a function that has stopped thinking.

Leaving the core out of the strategy. Extending and defending the core is a theme, and squeezing more from it is usually the largest single source of value and the cash for everything else. A strategy that is all about the new businesses has forgotten who pays.

Example: A services business's horizon-three item is an export pilot. It has had the same line in the plan for four years, with no milestones, no budget of its own and a sponsor who has since left. It is an orphan, not an option.

The five stages of the initiative portfolio

The shape of the initiative portfolio is one of the practices in the resourcing-and-commitment dimension of a stages-of-excellence strategy assessment. The five stages read:

• Lagging: Dozens of initiatives exist; nobody can list them, several contradict, none is clearly strategic.
• Basic: Initiatives are listed but unweighted, a flat backlog rather than a portfolio.
• Competent: A handful of initiatives each has an owner, a rationale and a link to a strategic theme.
• Advanced: The portfolio is balanced across horizons, with near-term initiatives funding longer-term horizons.
• Leading: The portfolio is actively managed: initiatives are killed, doubled or added as evidence arrives.

Most businesses are at the first or second stage. The move to the third is a cut, not a framework: the vital few, each with an owner and a reason. The horizons earn their place at the fourth.

Frequently asked questions

How many initiatives should a business have? A handful per strategic theme, and a handful of themes. If the list needs a spreadsheet, it is a backlog. The test is whether the top team can name them from memory and say what each is for.

Does this apply to a small business? Yes, in miniature. Horizon one is the business. Horizon two might be one new service or customer type being scaled. Horizon three might be a single experiment with a modest budget and a date. The discipline of naming which is which is the point, not the scale.

Is the three-horizons model out of date? The time assumption is. The purpose sort is not, and neither is the observation that each horizon needs its own money, measures and people.

How does this relate to a strategy on a page? The initiatives band of the page is where the sorted portfolio lands: the few things the business will do, each with its horizon, owner and first step.

What if all our initiatives are in horizon one? Then the sort has told you something the initiative list could not. It is not necessarily wrong for this year. The momentum case will show what it means for year five, and someone has to answer where the next core is coming from.

Is your initiative list a portfolio? Take the check. · The stages of excellence in strategy · Strategy as a hessian sack of initiatives