Scenario Assumption Traceability

A management team builds a downside cash scenario using lower sales, slower collection and higher supplier prices. Six weeks later, no one remembers which inputs came from signed contracts, which came from a sales forecast and which were rough judgments.

Scenario assumption traceability records that lineage so the model can be checked and updated. Umbrex discusses identifying and challenging critical assumptions in scenarios, and Model Reef describes version and approval tracking for scenario models, but these are management practices, not a required format for every spreadsheet.

List the scenario by defining the question and horizon before collecting inputs, because a short-term liquidity stress and a five-year growth case need different evidence. Identify each material input, such as price, volume, conversion, working-capital timing and funding cost, and prioritise sensitive assumptions.

Record the source, so that a contract, system report, market quote or expert estimate is linked to the assumption it supports, and mark confidence with plain labels rather than fake precision, since "signed agreement" and "management judgment" should not look equally certain. Keep the date, because a supplier quote from last year may be stale, and name the owner, since the team responsible for sales volumes may differ from the owner of tax timing or capacity costs.

Record transformations too: if weekly sales are converted to monthly cash receipts using a lag, preserve the calculation and lag rule. Separate facts from choices, since a current contract price is evidence while choosing to model a 10% downside shock is a scenario decision.

Preserve versions so a revised assumption does not overwrite the old one without a change log, and link to outputs so that when a liquidity result changes the reviewer sees which inputs changed and by how much. Check consistency and avoid double shocks, because a scenario with lower sales and unchanged variable costs per total period may be internally inconsistent, and a lower order forecast plus a separate "lost customer" adjustment may count the same event twice.

Use ranges, so an uncertain input can be represented by plausible bounds and a reason rather than one unjustified point estimate, and review dependencies, since currency, material price and customer demand may move together. Track approvals so a scenario shared with the board carries a date and reviewer record under company practice, keep access appropriate for sensitive pricing, salary and customer information, and inspect external rules because tax, employment and regulatory assumptions can change by country and need current local sources and specialist advice.

Watch copied cells and define units, as a spreadsheet can contain hardcoded numbers buried in formulas or hidden tabs, and "Revenue 10" could mean thousands, millions or a percentage change. Make challenge possible, so a reviewer can trace the lowest-cash month to its input evidence without asking the model author to reconstruct it, and update upon events, such as a signed customer renewal replacing an uncertain probability assumption.

Retain discarded cases, because a scenario no longer favoured can still explain why a decision was made earlier, and keep the register concise by ranking entries by sensitivity and risk so trivial notes do not bury the assumptions that move the decision. For an owner, traceability means every important scenario number has a provenance, calculation, owner and date, which makes disagreement useful rather than a fight over mysterious cells.