Speculative Design for Strategists: Scenarios Specific Enough to Argue With

Speculative design gives strategists a way to make scenarios concrete enough to argue with. Instead of four named worlds on a 2×2, it puts specific products, services and competitors inside each future, then works back from them to the choices in the current strategy: where to play, how to win, which capabilities to build. The output is not a forecast. It is a set of strategic choices tested against futures the plan never considered.

Most strategy functions already have a scenario set, presented at the start of the planning cycle and then reduced to context slides once the market model and the five-year plan take over. The scenarios were rarely wrong. They were too abstract to change a choice. For the general picture, see what speculative design is and our practitioner's workflow; this guide does not repeat them.

What strategists already bring to speculative design

Strategists start with more of this practice than most roles. The front half of the speculative design workflow is foresight, and corporate scenario planning is its best-known commercial form, with a founding story that belongs to strategy: the scenario work Pierre Wack led in Royal Dutch Shell's Group Planning department in the 1970s, written up in two Harvard Business Review articles in 1985, "Scenarios: Uncharted Waters Ahead" and "Scenarios: Shooting the Rapids".

Wack's central point matters here. Scenarios that merely set out alternative outcomes of obvious uncertainties do little for the people who have to decide. The real target is the decision-maker's own model of how the world works, what Wack called the microcosm. A scenario set that leaves the microcosm untouched has failed, however well researched. Speculative design is largely a set of techniques for not leaving it untouched.

You also have a language for choices. A. G. Lafley and Roger Martin's Playing to Win (2013) frames strategy as five linked choices: a winning aspiration, where to play, how to win, the capabilities required, and the management systems that support them. That cascade is where the work has to land. Market models, SWOT, competitor maps and the assumptions page complete the kit. If your organisation also has a foresight team, our guide for foresight practitioners covers their side.

The horizon problem: the five-year plan is always five years long

Strategy has a peculiar horizon problem. The plan's window, usually three to five years, rolls forward every year, so the far future never arrives in the document: the window moves with you. Each refresh extrapolates the same distance from a slightly different present, and anything maturing beyond the window becomes a risk, a trend slide or a footnote.

Scenarios are supposed to fix this, and on paper they do: most look ten or twenty years out. In practice they are read before the planning work starts and then set aside, because nothing in the plan is obliged to answer them. The long horizon exists in the strategy function; it is just not wired to any choice.

Speculative design works beyond ten years, but for a strategist the return journey matters more than the distance: every future has to come back as a challenge to a specific choice, inside the window the plan covers.

Where speculative design fits in the strategy process

The workflow does not sit beside the strategy process; it plugs into artefacts you already maintain.

Most teams skip the bottom row, yet it is what makes the rest usable: a backcast yields something a strategy document can hold, a list of events which, if they happen, reopen a choice.

A worked example: a snacks business and the appetite pill

Take a hypothetical strategy team at a European snacks and confectionery business. The company is invented; the signals are not.

The brief. The planning question is where the portfolio should play over the next five years. The speculative brief pairs a source of change with a domain and fixes the horizon: how would widespread use of appetite-suppressing medication change who buys snacks, and why, by 2037? One assumption the work may break is written down: that consumption per head is stable enough to model as a constant.

Signals. Dated events, not opinions about weight-loss drugs:

• •On 4 June 2021, Novo Nordisk announced US Food and Drug Administration approval of Wegovy, a once-weekly semaglutide injection, for chronic weight management.
• •On 6 May 2025, WeightWatchers filed for Chapter 11 protection, restructuring around the telehealth service it had used since 2023 to help members obtain prescriptions for these drugs.
• •The FDA approved two once-daily GLP-1 pills for weight management within four months: Novo Nordisk's Wegovy pill on 22 December 2025 and Eli Lilly's Foundayo on 1 April 2026.
• •Semaglutide's patent protection in India expired on 20 March 2026; within two days, generic versions from Sun Pharma, Dr Reddy's, Alkem and Zydus could be ordered from Indian pharmacies.

Filed one at a time, these look like healthcare news. Together they describe a path from injection to pill, from branded to generic, and from competing with the drugs to prescribing them.

The future wheel. The future wheel starts from the signal fully realised: appetite is a cheap, daily, prescribable setting, used by a substantial share of adults in the company's markets. The first ring is what the trade press already says: fewer calories, smaller portions, fewer impulse purchases. The second ring is where strategy lives. The population the market model treats as one curve becomes two, medicated and unmedicated eaters, and the prescriber becomes a channel, possibly a gatekeeper, between a food company and some of its customers. In the third ring, whoever pays for the medication takes an interest in what its patients eat, and food sold alongside a prescription attracts regulation food has never had.

The concept. Scenarios built from a wheel like this tend to come out as adjectives, such as "the medicated majority". The concept discipline stops that by naming something somebody will sell in 2037:

Measure — "Eat to your dose." A food programme sold through the telehealth app that prescribes the medication: small, high-protein meals matched to the user's current dose, delivered weekly, part-funded by the employer health plan that pays for the drug. The sketch is one phone screen: the user logs a dose increase, and the app offers next week's box.

Naming it forces the questions the scenario let the team avoid. Who runs Measure: a food company, a pharmacy chain or the prescriber? Who owns the customer, and the dose data? Is it food or medicine? The team's first real argument is whether their business is Measure's supplier, partner or casualty, and that argument is the point.

Implications. From human experience outwards: eating becomes something managed, and a family meal runs at two speeds. The treat, the category's emotional core, becomes a deliberate occasion rather than an everyday one. The business would be designing for people whose appetite is pharmacologically suppressed, using health data it does not hold today, and selling through a prescriber invites rules on health claims and marketing. The shipwreck: a programme that lives inside a prescriber's app can be switched off by the prescriber.

Backcasting. Backcasting from Measure looks for the rungs that had to be crossed. A threshold: generic oral medication cheap enough to be ordinary, and India shows how fast price moves once patents lapse. A scale of adoption: pills rather than injections. Institutional rungs: reimbursement decisions in the largest markets, a ruling on how food sold with a prescription is classed, data-sharing terms between prescribers and suppliers. Each is an event with a date and an owner, so each can be watched.

What changes this quarter. The output goes back into the cascade, in the cascade's own terms:

• •Winning aspiration: unchanged.
• •Where to play: the market model is rebuilt with two populations instead of one, making the medicated segment an explicit, revisable input.
• •How to win: the current answer, scale in impulse channels, is flagged as the most exposed choice.
• •Capabilities: reformulating two existing lines for protein and portion is funded, because it helps in this future and costs little in the others.
• •Management systems: the strategy document gains a page of dated triggers, such as a generic oral GLP-1 approved in the largest market, each with an owner and the choice it reopens.

Measure is never built. It has done its job if the plan now contains choices it would not otherwise have made.

Where speculative design goes wrong for strategists

Probability language. Strategists are trained to weight outcomes, and the habit leaks. Scenarios get labelled base case, upside and downside, or given percentages, and the organisation then plans for the base case alone. Speculative futures are plausible, not probable; nothing in the method produces a likelihood. The aim is preparedness across several futures rather than precision about one, the argument of NKD_6, against forecast-led strategy.

Falling in love with a scenario. The strategist who built the most vivid future becomes its advocate, and the backcast is run to justify a favoured investment rather than test it. Build at least two futures that make different choices look wrong, and treat any move that pays off in only one of them as a bet that needs its own case.

The scenario that lives beside the plan. The work arrives as a separate document, in its own vocabulary, after the planning calendar has closed: Wack's untouched microcosm in a newer format. Time it to the planning cycle and deliver it as amendments to the cascade, the market model and the assumptions page. If the sponsor is not in the room where the cascade is decided, send them our guide for leaders; absorption is usually theirs to solve.

Sizing the concept. How big is Measure? Any number will be extrapolated from the present, the very baseline the work exists to escape. Size the rungs instead: how many people self-pay at a given price, how long a reimbursement decision takes.

A first exercise for this week

You need your strategy document and two colleagues.

1. 1.List the assumptions built into the market model, not the ones on the risks page. Pick the one that would hurt most if it moved.
2. 2.Find three dated signals that suggest it could move: who did what, and when.
3. 3.Run a future wheel from the strongest signal, set at least ten years out, for forty-five minutes. Do not stop at the first ring.
4. 4.From the second or third ring, design a competitor that does not exist yet: name, tagline, one-paragraph description, a sketch of its home page.
5. 5.Write that competitor's cascade on one page: aspiration, where it plays, how it wins.
6. 6.Put your own five choices beside it and mark each as sound, exposed or wrong.
7. 7.For each exposed choice, write down the event that would tell you to act, and who would notice it first.

Where to learn speculative design as a strategist

Speculative Design Basics is our self-paced online course: 42 lessons, lifetime access, no prior experience assumed, taught by J. Paul Neeley. For strategists the Creating Value chapter matters most, because it covers how speculative design relates to foresight and strategy; the Tools & Methods chapters cover signals, future wheels and backcasting, and the Process chapter's treatment of time horizons answers the rolling five-year window. If you already have grounding, the live Speculative Design Masterclass runs over five sessions; Basics first is encouraged, not required. Other roles are covered in speculative design for every role.

Key takeaways

• •Strategists already own scenario planning, the front half of speculative design, and the language its results must land in: the strategy cascade.
• •Scenarios fail when they leave the decision-maker's model of the world untouched; named concepts inside them make them arguable.
• •The rolling planning window keeps the far future out of the plan. The backcast brings it inside.
• •Re-run the cascade from inside each future to see which choice breaks first; fund moves that help across futures before bets on one.
• •Speculative futures are plausible, not probable. Calling one the base case turns exploration back into forecasting.
• •The deliverable is an amended strategy with dated triggers and owners, not a scenario report.