Time Horizons: Why Speculative Design Works Best Beyond Ten Years

A time horizon is how far into the future a piece of work is set. In speculative design it is the first decision made and the one that governs everything after it, because it determines which of today's constraints are still binding on the conversation. We set it beyond ten years, and rarely past twenty-five. Nearer than ten, the future turns out to be the current plan with better rendering; further than twenty-five, people stop being able to picture the world or feel responsible for it.

This is the single largest practical difference between speculative design and the adjacent practices it gets confused with. Innovation work, product strategy and most design research operate inside the organisation's own planning window. Speculative design deliberately does not, and most of what is distinctive about the method follows from that one choice.

The horizons organisations actually run on

Most organisations think about two years ahead. This is not a failure of imagination; it is what their machinery is built to produce. Budget rounds allocate money a year at a time, reporting is quarterly, and long-term incentive plans typically pay out over three. Strategic plans are often written for five years, but years four and five are rarely funded and almost never staffed — they show direction rather than getting delivered. Two years is the distance at which a plan still has a named owner, a number attached, and consequences for the person holding it.

Software teams run much shorter. A team shipping continuously plans in two-to-six-week increments, with quarterly objectives above that. Anything past the current quarter is a slide rather than a commitment, and everyone in the room knows it.

None of this is a criticism. These cadences are how anything gets delivered, and an organisation that planned everything on a fifteen-year clock would ship nothing. The problem arises only when the same clock is applied to a different kind of question.

What a near horizon does to the conversation

Inside two years, every constraint that governs the organisation today is still live. The platform, the suppliers, the headcount, the procurement contracts, this year's budget owner, the current org chart — all of it is real, present in the room, and enforceable. Ideas get checked against those constraints before they have finished being spoken. The result is predictable: what comes back from a two-year "future" workshop is next year's roadmap with a nicer render.

There is a second, quieter problem. At two years you are mostly right, because the near future is largely determined by decisions already taken — capital committed, hires made, contracts signed. There is very little to discover, and discovery is the point. It is also the horizon at which linear intuition performs best, which is the flip side of Amara's Law: the errors that matter are not in the two-year estimate, they are in the ten-year one.

Ten years, and what expires by then

Ten years out, most of the things that were binding have stopped binding. The budget cycle has turned ten times. The technology stack has been replaced, probably twice. The procurement contract has been retendered. The market position currently being defended has either been lost or become uninteresting. A large share of the job titles in the room no longer exist in that form, including some held by the people arguing.

What survives is what makes ten years workable rather than merely distant: the sector still exists, the regulatory frame is recognisable, customers are still people with bodies and households, and the organisation can see itself in the picture. Ten years is far enough that nobody has to defend anything, and close enough that everyone still has a stake. That combination is narrow, and it is the whole reason for the band.

The real function of the horizon is permission

Here is the part that is usually left out of methodological accounts, and it is the most important thing on this page: the distant horizon does more social work than analytical work.

There is a sentence most people in an organisation cannot say. It is some version of the thing we are currently building is the wrong thing. Said about next year, that sentence is a criticism of a named colleague's plan, delivered in front of them, by someone with a career to consider. It is not a contribution; it is an incident. So the room does not say it, and the room's silence gets recorded as agreement.

Said about 2040, the same sentence is not about anybody. Nobody owns 2040. No one's promotion depends on it, no one wrote the deck, and no one has to walk out of the room having lost an argument. The distance removes authorship, and once authorship is removed the thing becomes sayable. In our experience this is where the genuinely useful material in a futures workshop comes from — not from novel information, which the organisation usually already has somewhere, but from what people already believed and had no safe way to voice.

It works in both directions. A distant horizon also gives people permission to be enthusiastic about something the current plan has no room for, which is equally hard to say out loud in a quarterly review.

The consequence is that the horizon is a facilitation decision as much as a methodological one. If a room is guarded — a regulated industry, a team six weeks after a restructure, three levels of hierarchy sitting at the same table — the answer is to push the horizon further out, not to encourage people harder. Guarded rooms need more distance, not more warmth.

The same property is also the method's easiest abuse. A distant horizon will happily absorb a live disagreement about next quarter: everyone says the difficult thing, everyone feels heard, and nothing ever comes back to the present. That is precisely why the work is not finished at the scenario, and why backcasting is not an optional final stage.

Where the far end goes wrong

Past roughly twenty-five years, engagement drops, and it drops for three separate reasons.

Responsibility evaporates. Nobody in the room expects to be accountable for 2075. The same distance that made criticism sayable now makes it inconsequential. Permission has tipped into indifference.

World-building overwhelms the argument. The further out you go, the more of the world has to be rebuilt before any single concept can be understood at all. The ratio goes wrong: the session is spent explaining the context rather than arguing about the thing sitting inside it.

Genre takes over. At long range the material drifts towards inherited science-fiction imagery, and audiences recognise the aesthetic instantly. Work that reads as genre gets admired, photographed, and filed.

This band is an observation from practice, not a law, and it moves with the audience. Foresters, hydrologists, climate scientists, actuaries and anyone working on nuclear waste hold fifty- or hundred-year horizons natively, because their profession trains it. With those rooms the constraint sits much further out, and the ten-year default would be far too near.

Choosing a horizon for a brief

Fix the horizon in writing before any other work begins, because drift is always inward. Left unfixed, a room will migrate steadily towards the near end of the range, one reasonable-sounding step at a time. Three principles do most of the work.

Go at least one full replacement cycle past the slowest thing your answer depends on. Consumer software replaces itself every two or three years, so ten years is already several complete generations of platform, interface and business model — ample room. Health systems, rail, grid, water and housing run on asset lives of thirty to sixty years, procurement rounds measured in years, and evidence and approval cycles measured in a decade. For those, ten years is barely one turn of the wheel, and a ten-year health brief tends to produce a service redesign rather than a future. Push them out.

Go past the organisation's longest existing plan. Find the furthest-dated document anybody has and set the horizon beyond it. If there is a 2035 net-zero commitment, then 2035 is not a horizon — it is a deliverable, with an owner who will defend it.

Calibrate to the room, not only to the sector. Two people from the same company can have very different native horizons; a treasury function and a growth team will not stretch equally. And when the choice is genuinely unclear, go further and come back. A room that has spent a morning arguing about 2050 can discuss 2035 without flinching. The reverse almost never happens.

Where to take this next

Make the horizon an explicit, written decision rather than an assumption, and put it further out than feels comfortable. Almost every weak futures project we see is traceable to a horizon that was never really chosen.

Our Speculative Design Basics course covers how that decision runs through the rest of the process — briefs, signals, scenarios and the concepts that come out of them — at your own pace.

Key takeaways

• •Most organisations plan on about two years because their budget, reporting and incentive systems are built to that length; many software teams plan in two-to-six-week increments.
• •Inside that window every current constraint is still enforceable, so "future" concepts come back as next year's roadmap with better rendering.
• •Beyond ten years, budgets, contracts, technology stacks, market positions and most job titles have expired, while the sector, the regulator and the customers remain recognisable — that combination is what makes the horizon productive.
• •The horizon's main function is social, not analytical: distance removes authorship, which lets people say things about the plan that they cannot say about next year without it reading as criticism.
• •Past roughly twenty-five years, responsibility evaporates, world-building swamps the argument, and the work reads as genre — though specialist audiences hold much longer horizons natively.
• •Choose the horizon by the replacement cycle of the slowest thing involved, set it beyond the organisation's longest existing plan, and fix it in writing, because unfixed horizons always drift nearer.