Strategy under uncertainty is the problem of choosing when the diagnosis will not settle. Hugh Courtney, Jane Kirkland and Patrick Viguerie gave that problem a usable shape in Harvard Business Review in 1997. Their argument starts from what happens when the world is treated as either knowable or unknowable. A manager is then left with two poor options, which are a confident plan built on a forecast nobody believes and a retreat into instinct. Between those two sits everything that remains once all the knowable analysis has run, and the three authors called that residual uncertainty. Some assumptions on a strategy's list admit no test at any price, and residual uncertainty names exactly what blocks those tests.
The sections below give the four levels of residual uncertainty and the three postures a team can take towards any of them. A table then sets the levels against the posture and the tool. The last two sections cover scenario planning for what no forecast reaches and real options for holding a choice open at a price.
The four levels of residual uncertainty
Courtney, Kirkland and Viguerie sorted what remains after analysis into four levels, and the level decides which tools can answer.
- A clear enough future. One forecast survives the analysis, narrow enough to plan against.
- Alternate futures. A small number of discrete outcomes, one of which will happen, with no way to say in advance which one.
- A range of futures. A bounded range holding no discrete outcomes, so the answer could land anywhere between two limits.
- True ambiguity. No bound on what counts as plausible, because the market, the technology and the rules are all moving at once.
Level four is rare and it is temporary. A market sits there while something new is arriving, and it drops to level three or level two as the first facts land. A team's level is therefore a reading of a moment, and never a permanent property of the industry.
The three strategic postures
A level says what can be known. A posture says what the organisation does about it, and the three authors name three postures.
Shaping the future means acting to make one outcome the one that happens, by setting a standard, building a market or committing at a scale that changes what competitors find worth doing.
Adapting to the future means taking the industry structure as it comes and competing on how fast the organisation notices a change. The aim is to be first to see which way things went and first to respond.
Reserving the right to play means buying a position that costs little now and keeps a larger move available later, such as a small stake, a pilot or a capability held ready.
Each posture rests on three kinds of move. A big bet is a large commitment that pays well in one set of outcomes and loses in the rest. An option costs little, pays well in a few outcomes and limits the loss everywhere else. A no regrets move pays whatever happens, which is why the authors recommend it at every level, including level one.
Matching the level to a posture and a tool
The levels and the postures meet in one table, and the table is what a team actually works from.
| Level | What remains after analysis | Posture the level affords | The tool that answers |
|---|---|---|---|
| One, a clear enough future | A single forecast narrow enough to plan against | Any of the three, with shaping a question of scale | Market segmentation, competitor cost analysis, the five forces and discounted cash flow |
| Two, alternate futures | A short list of discrete outcomes | Shaping, or reserving the right to play until one outcome shows | Decision analysis, option valuation models and game theory |
| Three, a range of futures | A bounded range holding nothing discrete | Adapting, or reserving the right to play | Latent demand research, technology forecasting and scenario planning |
| Four, true ambiguity | No bound on what counts as plausible | Shaping, or reserving the right to play cheaply | Analogies with comparable markets and pattern recognition |
The posture column names what a level usually affords, and a team stays free to take any of the three. Courtney, Kirkland and Viguerie put all three postures on the table at every level. A company with the resources to set a standard can shape a level four market, and a company without them will struggle to shape a level two one. What the level settles is the tool. A discounted cash flow run on a level four market produces a number that carries the authority of a guess and the persuasive power of a spreadsheet, and that is the error the four levels exist to prevent.
Scenario planning for what no forecast reaches
Scenario planning is the tool the middle two levels share, and it does a job no forecast does. A forecast produces one number with error bars around it. A set of scenarios produces several descriptions of how the world could look, each one consistent with itself and detailed enough for a team to ask what the strategy would do inside it.
Building that set of scenarios is where the discipline sits. Scenarios have to be few enough to work with, which usually means three or four. Each one has to lead somewhere different from the others. Together they have to be wide enough to hold the range the team believes is possible. If a set of scenarios leaves the strategy doing the same thing in every one of them, it has failed, because the fortnight it cost bought no decision.
Real options and the price of keeping a choice open
Scenarios describe the futures a team cannot choose between. Reserving the right to play is the posture for exactly that position, and a real option is how a team pays for it. A real option is a small commitment made now that buys the right to make a larger one later. Options of this kind take several forms.
- A pilot in a second market, which costs a quarter and buys knowledge of whether the market responds at all.
- A licence on a technology the company has no plan to ship this year.
- A minority investment in a supplier whose part may become the constraint.
- An architecture built to accept a component nobody has chosen yet.
Each of those costs something today, and none of them commits the organisation to the larger move. What makes the idea useful to a product team is the pricing. An option is worth taking when the cost of holding it is small against the loss the organisation would take if the larger move were no longer available. If a team cannot state both numbers, it has not finished the argument.
Watching a bet after it is placed
Every posture on the table commits the organisation to something before the uncertainty clears. Shaping spends the most and adapting spends the least, and both of them rest on the same later step. That step is noticing what actually happened. A posture chosen in January and reviewed in December has been an act of faith for eleven months.
Noticing takes a small set of measures, chosen in advance and built from the strategy's own assumptions, so that a number moving says something specific about the bet. Choosing those measures is a piece of work in its own right. If they are chosen badly, the team ends up with a dashboard that reports activity while the strategy quietly fails.