How Many Scenarios Should You Model? A Practical Guide to Base, Upside, Downside and Stress Cases

## The short answer

For most planning exercises, three to four scenarios is the right number: a base case, an upside, a downside, and optionally a stress case. Fewer than three and you are barely planning for uncertainty; more than five and the analysis becomes unwieldy, the scenarios start to blur together, and decision-makers stop engaging. The goal is not to cover every possibility but to bracket the range of outcomes that would change what you decide.

## Why three to four works

The number isn't arbitrary. It reflects how decisions actually get made. Each scenario you add must earn its place by changing a decision; if two scenarios lead to the same actions, you only needed one. Three to four scenarios typically span the space where decisions flip - where you would hire or freeze, invest or hold, expand or consolidate - without drowning the audience in detail.

There is also a cognitive limit. People can hold a handful of distinct futures in mind and reason about trade-offs between them. Push past that and the exercise collapses into a spreadsheet nobody internalises.

## The four scenarios that cover most cases

### 1. Base case

Your single most likely path, built on best-estimate assumptions. This is the reference point everything else is measured against, and it usually doubles as your budget or plan of record. It should be defensible, not optimistic - the most common failure is a "base" case that is quietly an aspiration.

### 2. Upside case

What happens if things go better than expected: stronger demand, better pricing, faster conversion. The upside matters more than people assume. If you only plan for downside, you under-invest in capacity and miss the moment when growth arrives. The upside case answers "could we even cope with success, and what would we need in place first?"

### 3. Downside case

A realistic adverse path: softer demand, margin compression, a delayed launch, a key customer churning. This is where most of the protective value sits. The downside case tests whether your cost base, cash runway and commitments survive a bad-but-plausible run of events.

### 4. Stress case (optional but powerful)

A low-probability, high-impact combination - several bad things at once. The point of the stress case is not realism; it is to find the breaking point. At what level of revenue shortfall does cash run out? When does a covenant breach? Knowing the cliff edge is more useful than pretending you'll never approach it.

## When to use fewer or more

**Use two scenarios** only for quick, low-stakes decisions where a simple better/worse contrast is enough to frame the choice.

**Use five or more** rarely, and only when distinct, named external futures genuinely demand it - for example, modelling separate regulatory regimes or clearly divergent market structures. Even then, group them so the audience can reason in pairs.

## How to keep scenarios distinct

The most common mistake is producing scenarios that are just the base case scaled up and down by a fixed percentage. That tells you about sensitivity, not about different futures. To keep scenarios meaningfully separate:

- **Change the narrative, not just the numbers.** Each scenario should have a one-line story a colleague could repeat from memory.

- **Move multiple drivers together, coherently.** In a downturn, volume falls *and* pricing weakens *and* bad debt rises - they correlate, so model them together.

- **Avoid overlapping ranges.** If the downside's best outcome overlaps the base case's worst, the boundary is fuzzy and the comparison loses force.

- **Assign rough probabilities** to keep the conversation honest about how seriously to treat each one.

## Tie scenarios to triggers

Scenarios are only half the work. The other half is defining the observable signals that tell you which one is unfolding - leading indicators like enquiry volume, pipeline velocity, supplier lead times or input costs. When a trigger fires, you already know the playbook because you modelled it in advance. This turns scenario planning from an annual ritual into a live decision aid.

This is precisely the layer enterprise tooling adds, and an area where neart.ai builds enterprise-grade products: maintaining several scenarios off one driver model and linking them to real signals so the plan updates as evidence arrives, rather than gathering dust between budget cycles.

## A quick self-check

Before you finalise your set, ask:

- Does each scenario lead to a *different* decision? If not, merge or cut it.

- Could I summarise each in one sentence? If not, it isn't distinct enough.

- Do I know what signal would tell me this scenario is happening? If not, add a trigger.

## Practical takeaway

Start with three - base, upside, downside - and add a stress case when the downside risk is severe or irreversible. Resist the urge to model more. Every scenario should change a decision, carry a distinct one-line story, move several drivers together, and come with a trigger that tells you when it's arriving. Quality and distinctness beat quantity every time.