Skip to content
LeanSimCorpIndustry problems solved
Explore site
Illustrative production planners reviewing demand and schedule scenarios above a manufacturing floor

Scenario &
what-if comparison.

Compare the options on one basis.

The value is not in any single run but in the comparison: same model, same demand, same assumptions, one thing changed at a time.

AI-generated illustrative operation

The real challenges.

Two options compared on different assumptions tell you nothing about either. Explore five things a scenario study has to control, using the cards or arrows below.

How simulation helps

A common basis

One change at a time

Difference or noise

Robustness

Presenting the choice

Challenge 1 of 5

A common basis

Hold demand, assumptions and measures identical across every scenario.

  • Fix the common inputs
  • Agree the measures first
  • Document the basis once

One change at a time

Change one thing per scenario so a difference can be attributed.

  • Define each scenario precisely
  • Isolate the variable
  • Combine only deliberately

Difference or noise

Separate a real difference from the variation of the model itself.

  • Replicate every scenario
  • Compare across replications
  • Report the uncertainty

Robustness

Check whether the ranking survives demand and input uncertainty.

  • Vary the uncertain inputs
  • Look for options that hold
  • Flag fragile winners

Presenting the choice

Report the comparison in terms of the decision being made.

  • Report against the measures agreed
  • Show what separates the options
  • State what would change the answer
AI-generated illustrative planning simulation linking demand ranges, product batches, production cells, and finished stock; not a partner model

Explore all five challenges · Select a card or use the arrows

The results.

A scenario study is only as good as its controls. What it produces is a defensible comparison, and a clear statement of what would change the answer.

One basis for every option

Demand, assumptions and measures held identical, so differences come from the change rather than the setup.

Attributable differences

One variable moved per scenario, so a result can be traced to a cause.

Noise separated out

Replicated runs, so a difference is distinguished from the model’s own variation.

Options tested for robustness

Whether a ranking survives demand and input uncertainty, not only the central case.

A comparison you can decide from

Results reported against the measures the decision is actually made on.

What you take away

Scenario definitions and basis + Replicated comparison results + Robustness and sensitivity findings

Scenario & what-if comparison consulting

We define the options, hold everything else constant, and run each one enough times to separate a genuine difference from run-to-run variation. The output is a comparison that can be defended, with the conditions it depends on stated.

Explore our simulation project support

Fixing the common basis

Demand, calendar, assumptions and measures are fixed once and used by every scenario. Comparisons that drift — a different demand profile here, an extra shift there — produce differences that cannot be attributed to anything, which is the most common reason a scenario study fails to settle an argument.

One variable per scenario

Each scenario changes one thing. Where a combination is genuinely the option under consideration we model it as such, but deliberately and separately, so it is never unclear which change produced which part of the result.

Telling a difference from noise

A simulation model varies between runs. We replicate each scenario with different random seeds and compare across replications, so a small gap that falls inside normal variation is reported as inconclusive rather than as a winner.

Robustness, not just the central case

An option that is best only under one forecast is a risk, not a recommendation. We vary the uncertain inputs across a plausible range and report which options hold their position and which depend on conditions that may not occur.

Inputs for a credible comparison

Typical inputs are a validated model of the current operation, the options to be compared, the demand or forecast range to test against, and the measures the decision will be made on. We agree those measures before any scenario is run.

Reporting for the decision

Results are presented against the agreed measures, showing what separates the options, how confident the separation is, and what would change the answer. A ranked table without that context invites a decision the study does not actually support.

Compare the options
on a basis you can defend.

With expert simulation consultation.

Tell us the options you are weighing, the measures the decision rests on, and the uncertainty you need to test against. We’ll set out what a scenario study should cover.

More in Data & Analytics