---
title: "Scenario planning: replace one forecast with a few stories"
description: "When a key driver can go several ways, one forecast can sink an investment. Six steps of scenario planning, a worked hypothetical example and its limits."
url: "https://staging.khoi.co.uk/en/blog/lap-kich-ban-cho-chu-doanh-nghiep"
lang: en-GB
author: "Lê Quốc Khôi"
date: 2026-11-02
updated: 2026-11-02
topics: ["Strategy & leadership"]
tags: ["Scenario planning", "Strategic planning", "Uncertainty", "Investment decisions"]
---

# One number or several scenarios? Scenario planning for business owners

## Key takeaways

- A single forecast squeezes all uncertainty into one assumption. The public introduction to Wack's (1985) article notes that forecasting errors have become more frequent since the early 1970s, and are sometimes dramatically large (Wack, 1985).
- Scenario planning builds a few coherent stories about the future, then tests a decision against each one. Schoemaker (1995) presents it as a way to offset two familiar errors in decision making: overconfidence and tunnel vision (Schoemaker, 1995).
- "Good, middling, bad" is not scenario planning. Scenarios should differ in structure, not only in degree.
- In this post's hypothetical example (every figure is invented for illustration), buying the whole production line at once clears break-even in 1 of 4 scenarios, while a two-stage purchase clears it in 3 of 4. The price is some lost revenue if demand recovers fast.
- Each scenario needs a measurable early signal, tied to an action agreed in advance.

A furniture exporter has until the end of the fourth quarter to file next year's plan with its bank, together with a request to borrow for a second production line. The form has one box for expected revenue, and the box holds one number. The firm sells most of its output into a single market where both demand and policy are unsettled. The firm is hypothetical. [Scenario planning](https://staging.khoi.co.uk/en/framework/scenario-planning) keeps both unsettled questions on the table instead of squeezing them into that box.

## Why can one forecast spoil an investment decision?

Because the number hides every assumption behind it, and one wrong assumption drags the whole plan off course. Wack (1985) describes traditional planning as forecast-based, and says it worked reasonably well in the relatively stable 1950s and 1960s. From the early 1970s, forecasting errors came more often and occasionally on a dramatic scale. According to the article, most US companies kept using a range of forecasting techniques. The reason given: nobody had apparently found a better way to deal with economic uncertainty (Wack, 1985).

A single figure becomes dangerous in a big decision for three reasons.

- It becomes an anchor. Once it is in the application, every later discussion is about nudging it up or down a few points ([anchoring bias](https://staging.khoi.co.uk/en/glossary#anchoring-bias)).
- It feels certain. Schoemaker (1995) says scenarios help to offset two common errors in decision making, [overconfidence](https://staging.khoi.co.uk/en/glossary#overconfidence) and tunnel vision (Schoemaker, 1995).
- For inputs such as an importing country's trade policy, nobody may know even the probabilities of the outcomes ([Knightian uncertainty](https://staging.khoi.co.uk/en/glossary#knightian-uncertainty)). Writing down one figure then amounts to pretending to know.

## What is scenario planning, and how does it differ from best, base and worst cases?

Scenario planning means building a few coherent stories about the future around the forces that are both important and hard to predict, then asking how your strategy fares in each. Schoemaker (1995) describes it as a tool that captures a whole range of possibilities in rich detail, by identifying basic trends and uncertainties (Schoemaker, 1995). The six steps below follow the [Scenario planning](https://staging.khoi.co.uk/en/framework/scenario-planning) framework.

> **Figure:** [Scenario planning](https://staging.khoi.co.uk/en/framework/scenario-planning) The six steps of scenario planning, following the framework in the site library.

The difference from best, base and worst cases lies in structure. Cases of that kind usually push one variable up or down. Real scenarios differ in structure, not only in degree, and the framework's own notes flag this as the mistake to avoid.

## How do the six steps apply to a hypothetical furniture exporter?

A hypothetical example: the firm has revenue of VND 200 billion, 70 per cent of it sold into one market. A new line costs VND 40 billion, the bank would lend VND 24 billion, and at full capacity the line adds VND 60 billion of revenue a year. Every number in this section is invented by me for illustration. None belongs to a real company.

1. Focal question: should the firm borrow VND 24 billion for a second line over the next two years?
2. Driving forces: demand from the main buyers, new duties or standards in that market, the exchange rate, timber prices, freight rates. [PESTLE analysis](https://staging.khoi.co.uk/en/framework/pestel) is one way to sweep each group of factors so nothing is missed.
3. Rank by impact and uncertainty: suppose the firm can adjust its selling prices and timber contracts to soften exchange-rate and timber shocks. The two forces left, both large and hard to call, are demand from the main buyers and whether the door to that market narrows.
4. Build the scenarios: two axes with two ends each give four stories.
5. Stress-test the strategy: the table in the next section.
6. Monitor signposts: the list straight after the table.

> **Figure:** (matrix) The four scenarios for the hypothetical furniture exporter. All content is assumed, for illustration.

```json
{
  "type": "matrix",
  "x": { "label": "Demand from main buyers", "low": "Flat", "high": "Recovering" },
  "y": { "label": "Access to the main market", "low": "New barrier", "high": "Still open" },
  "cells": [
    { "key": "wait", "title": "Waiting game", "body": "Buyers hold steady and order no more. The door stays open." },
    { "key": "boom", "title": "Smooth expansion", "body": "Buyers order more and the door stays open." },
    { "key": "squeeze", "title": "Squeezed twice", "body": "Orders do not grow, and a new barrier arrives.", "risk": "high" },
    { "key": "narrow", "title": "Demand, narrow door", "body": "Buyers want more, but the barrier makes orders hard to deliver." }
  ]
}
```

Each cell is a different kind of story. In "Demand, narrow door" the firm does not lack customers. It lacks a route to them. The response (another market, another product) is nothing like the response in "Waiting game", where the orders have simply not come.

## How does scenario planning change the investment decision?

It shows which option survives in many scenarios and which lives in only one. Suppose the loan application carries a single figure: VND 45 billion of extra revenue, which means the line runs at 75 per cent of capacity, comfortably above an assumed break-even of 60 per cent. The plan looks safe. But 45 does not match any scenario, and it is higher than three of the four outcomes in the table below.

The table puts two options into each scenario: buy the whole line at once (capacity VND 60 billion), or buy in two stages (VND 20 billion now for capacity of VND 30 billion, the second stage only when the signals allow). The "extra revenue" column is assumed.

| Scenario | Extra revenue (VND bn) | One purchase: utilisation | Two stages: first-stage utilisation |
|:--|--:|--:|--:|
| Recovery, door open | 54 | 90% (clears) | 100% (clears) |
| Recovery, narrow door | 33 | 55% (short) | 100% (clears) |
| Flat, door open | 30 | 50% (short) | 100% (clears) |
| Flat, narrow door | 15 | 25% (short) | 50% (short) |

On these invented numbers, buying the whole line clears the 60 per cent bar in one scenario. Two stages clear it in three, and in the worst case only half the capital sits in machinery (VND 20 billion against 40). The price is in the first row: if demand recovers and the door stays open, the firm forgoes up to VND 24 billion of revenue a year until stage two is running (54 less 30). Staging fits if a missed opportunity costs less than stranded capital.

Signposts for the hypothetical firm:

- confirmed orders for the next two quarters against the capacity of the existing line: at 85 per cent or more for two quarters in a row, treat demand as recovering and start stage two;
- an official notice of new duties or standards in the main market: once a draft appears, treat the door as narrowing;
- the share of revenue from outside the main market: if it has not moved after two quarters, the search for new buyers is running late.

These thresholds are assumptions too. Each signal should be tied to an action agreed in advance, so the decision is not reopened every quarter.

## When does scenario planning fit, and when does it not?

It fits when the decision is long-lived and hard to reverse, and the outcome rests on one or two forces that could go several ways. That is the question the framework poses: before a long-term investment in a volatile industry, which futures should the strategy be tested against.

It does not fit small or reversible decisions. A forecast refreshed on a regular cycle, such as a [Rolling forecast cycle](https://staging.khoi.co.uk/en/framework/rolling-forecast-cycle), is cheaper and good enough there. For choosing a forecasting method, see [Choosing a forecasting method](https://staging.khoi.co.uk/en/framework/forecasting-method-selection).

Two limits deserve plain words.

First, scenarios are not forecasts. They do not tell you what will happen. They tell you what your decision can withstand.

Second, a scenario set only covers what the team that built it thought of. A shock nobody imagined appears in no cell. The public abstracts of the two source articles do not discuss this. It is my inference from how the method works, not a conclusion of Wack or Schoemaker. One way to reduce it, again by inference, is to have someone outside the team read the list of driving forces before the two axes are chosen.

## Going deeper: where does scenario planning come from, and what is missing?

Both articles sit behind a paywall. What follows rests on the publishers' public abstracts and on the framework's notes in the site library, not on the full texts.

Wack (1985) appeared in the September issue of Harvard Business Review. The public introduction to the article is about the limits of forecast-based planning since the early 1970s (Wack, 1985). The rest sits behind the paywall, and this post does not draw on it.

Schoemaker (1995) appeared in Sloan Management Review. According to its abstract, it describes how to build scenarios step by step through two case studies, Interpublic (an international advertising agency) and Anglo-American Corporation in South Africa, and how to use the resulting stories to plan a company's future (Schoemaker, 1995).

On criticism, neither public abstract offers quantitative evidence that firms using scenarios decide better. Treat the method as a thinking tool, and judge it by how decisions change when a signal appears, not by which scenario turned out to be "right".

For further reading in the library, [Macro-to-micro causal chain](https://staging.khoi.co.uk/en/framework/macro-to-micro-causal-chain) traces a macro driver down to a small business, and [Strategic management process](https://staging.khoi.co.uk/en/framework/strategic-management-process) places scenarios inside the wider strategy process.

## Frequently asked questions

### How is scenario planning different from sensitivity analysis?

Sensitivity analysis changes one assumption at a time and holds the rest still, to see which variable the result is most sensitive to. A scenario changes a whole set of assumptions that fit together, because in real life demand, policy and prices tend to move at once. The two work well together, with sensitivity checks used on the numbers inside each scenario.

### Do you have to put a probability on each scenario?

The site's scenario-planning framework treats each scenario as a coherent story, not a forecast, and its six steps include no probability-assignment stage. If the leadership team wants to assign probabilities anyway, label them as the team's own judgement rather than a measurement, and revisit them when the early signals arrive.

### Does a small business need scenario planning?

It does when the decision has two features. The money is hard to get back once spent, and the result hangs on one or two variables you do not control. For smaller decisions that can be reversed, a regularly updated forecast is enough and far cheaper.

## References

- Schoemaker, P.J.H. (1995) ‘Scenario planning: a tool for strategic thinking’, *Sloan Management Review*, 36(2), pp. 25–40. Available at: https://shop.sloanreview.mit.edu/store/scenario-planning-a-tool-for-strategic-thinking.
- Wack, P. (1985) ‘Scenarios: uncharted waters ahead’, *Harvard Business Review*, 63(5), pp. 73–89. Available at: https://hbr.org/1985/09/scenarios-uncharted-waters-ahead.

---

Canonical HTML version: https://staging.khoi.co.uk/en/blog/lap-kich-ban-cho-chu-doanh-nghiep