Scenario Analysis: How to Build Cases and Read the Range
Scenario analysis tests a plan by building a small number of complete alternative worlds, usually a downside, a base and an upside, and working every linked assumption through each world together. A fall in orders also delays collections and forces discounts, in the same case. The output is a range of outcomes, and the range is the finding; the average of the scenarios describes a world that cannot happen.
The base case gives an honest centre. The downside and the upside are built outward from that centre. The craft is consistency inside each world, and the payoff is a range that can be planned against instead of a single number that can only be hoped about. Moving one number while the rest stand still is a different tool, covered separately under sensitivity analysis.
What is a scenario, and why a whole world rather than one number?
Anyone who has planned an outdoor wedding has done scenario analysis under pressure. Rain causes all of them together, so the rain scenario is never just "the lawn gets wet": it is the backup hall, the covered walkway for catering, the photographer moving indoors, and the guest count that shrinks, all in the same evening. A scenario is one internally consistent world, every linked assumption moved together by a common cause.
Now the business version. A demand fall at Tessora Weaves does not arrive alone: the buyers who cut orders also pay slower, and the market that shrank also hardens prices. A case that cuts orders 30 per cent while collections and pricing stand untouched describes no possible world. The case moved one number and froze the consequences of the cause it named.
A downside case cuts orders 30 per cent and leaves collection days and pricing untouched. What is wrong with it?
How is a downside built so it is internally consistent?
Start from a cause, not a number, then trace the cause through every line it touches. Tessora Weaves' downside starts from one named event: US and EU retail demand falls. Trace it: Meridian Retail Group cuts orders 30 per cent; revenue drops to Rs 39,36,00,000; the same weak market stretches collections and forces season-end discounts; contribution halves through operating leverage; profit before tax (PBT) lands at Rs 58,00,000, a fall of nearly 90 per cent from a 30 per cent order cut. Every step is the consequence of the one before, which is what makes the case a world rather than a guess.
Before seeing the arithmetic: the downside cuts the largest buyer's orders by 30 per cent. Does PBT fall by more than, less than, or about 30 per cent?
How is the upside built, and why does optimism get checked too?
Consistency is not a pessimist's rule, so the upside is built with exactly the same discipline. Tessora Weaves' upside starts from its own named cause: a new buyer programme lifts revenue 15 per cent to Rs 55,20,00,000. Trace the consequences honestly: more volume needs more yarn bought at market prices, collections grow with sales, and the extra production strains the same fixed capacity. PBT works out to Rs 8,50,00,000. An upside whose only traced consequence is more profit is a hope case that borrowed the scenario format.
An upside case adds 15 per cent revenue and changes nothing else: no extra yarn purchases, no extra working capital. What is it really?
Why three cases, and not thirty?
Because each case must be argued as a world, and arguing takes work no spreadsheet can fake. A downside earns its place by having a named cause, a traced chain and a defensible severity; so does the upside. Thirty cases means thirty arguments. In practice that means none: thirty rows of arithmetic wearing case names. Few worlds, fully argued, beat many worlds asserted, and three is usually enough: the honest centre and one hard look in each direction.
A planning pack contains 25 scenarios. What has most likely happened?
What does the range show that no single case can?
Look at the three PBT outcomes on one line: Rs 58,00,000 to Rs 4,90,00,000 to Rs 8,50,00,000. The width of that span is the finding. It says: this plan lives in a world where profit can be almost nothing or almost double, driven mostly by one demand factor, and any commitment, a loan instalment, a hiring plan, a new unit, must be survivable across the whole span, not just at the centre. A single number, however honest, could never have said that.
A loan instalment plan is comfortable at the base and the upside, and impossible in the downside. What has the range just shown?
Why is the average of the scenarios a trap?
Averaging the three PBTs gives roughly Rs 4,66,00,000, a tidy, plausible-looking number that no world on the list produces. The downside world cannot fund a plan sized to it, and the upside world overfunds it. Averaging threw away the range, and the range was the entire finding. Scenario analysis compressed to its mean is the one summary that betrays it. If a single number is truly needed, report the base, and staple the range to it.
A colleague reports the analysis upward as one number: the average of the three cases. What information just disappeared?
How does the range turn into real decisions?
A range nobody acts on is decoration, so here is the working routine. First, commitment testing: every fixed promise, a loan instalment, a lease, a hiring plan, is laid against the downside column, and a promise the downside cannot carry gets resized before signing, not renegotiated after. Tessora Weaves' downside PBT of Rs 58,00,000 says plainly: do not sign fixed obligations above roughly that level without a buffer standing behind them.
Second, trigger points: the downside's early signposts become pre-agreed actions. The decision is made in calm weather. If Meridian Retail Group's orders run 10 per cent behind for two consecutive months, hiring freezes and the discretionary spend stops, automatically. The wedding planner's version: if the forecast turns by Tuesday, the backup hall gets booked, no debate required on the day. Deciding the response before the world picks a column is the entire practical payoff of having built the columns.
Why agree the trigger actions in advance, rather than deciding when the orders actually fall?
The error that gets made, and what it costs
The planner who sizes the funding plan to the averaged PBT of Rs 4,66,00,000. The averaged figure feels prudent, sits between the extremes, and fits precisely the one world that cannot occur. When the downside arrives, the plan cannot be funded; when the upside arrives, the money sits idle. The plan was built for the mean of three worlds instead of being tested against each of them.
The cost is a plan that is wrong in every world on its own list, purchased at the price of looking balanced.
Which single sentence best describes the deliverable of a good scenario analysis?
References
| Source | Document | Where |
|---|---|---|
| Reserve Bank of India (RBI) | Published risk material where scenario analysis is used | rbi.org.in |
Tessora Weaves Private Limited and Meridian Retail Group are invented.
Educational material. Not advice on any investment, tax, budget or market position.
