Loading
Economics · Scenario & Stress Testing
Know which assumptions can change before the decision stops working.
For boards and executives who distrust optimistic base cases but resist arbitrary pessimism. We test which assumptions actually matter, build coherent downside and severe-but-plausible scenarios, and define the failure thresholds and triggers that tell management to proceed, pause, adapt or stop.
- Sensitivity · Scenario · Stress · Reverse stress
- Failure thresholds made explicit
- Triggers with owners and actions
Trusted by established businesses
The constraint
When this work earns its place
Stress testing is not for every decision. It is for the ones where being wrong about an assumption is expensive, irreversible or both.
A major investment is about to be approved
The business case looks sound, the base case stacks up, and nobody has established which of its assumptions could quietly sink it after the money is committed.
A growth commitment locks in cost ahead of demand
Hiring, capacity, inventory or marketing spend is being committed on a demand projection. If demand arrives late or light, the cost stays and the revenue doesn't.
A forecast nobody fully trusts
The plan rests on a forecast that feels optimistic to half the room and reasonable to the other half, and the disagreement has never been resolved with analysis.
A price or business-model change
Repricing, repackaging or changing how the business monetises shifts several variables at once (volume, margin, churn, competitive response) and intuition can't hold them all.
Volatile costs or demand
Input costs, exchange rates or demand have become genuinely volatile, and the annual plan's single set of numbers no longer describes the range of futures the business faces.
The board is challenging the base case
Directors are asking 'what would have to be true?' and 'what if it isn't?', questions the current model was never built to answer.
A recovery plan must survive its downside
A turnaround or recovery plan is being committed to. If it only works in its own base case, it isn't a plan. It's a second bet.
A base case that has never been stressed is not a plan. It is a hope with a spreadsheet attached. The cost of testing it is always smaller than the cost of discovering its breaking point live.
The approach
Severe but plausible: not arbitrary pessimism
The failure mode of most downside analysis is arbitrariness: someone shaves twenty per cent off revenue, calls it the downside case, and everyone privately ignores it. We work differently. First we map the assumptions and dependencies: the model's logic, the variables that move together, the external drivers, the operational constraints and the criteria the decision will actually be judged against. Then we test the model four complementary ways, define explicitly what failure means for this decision, and design the responses in advance: which indicator to watch, what level triggers action, who owns it, and what they do. The output is not a more pessimistic forecast. It is knowing which assumptions matter, where the breakpoints sit, and what management does when reality starts drifting toward one.
What we analyse
Four ways of testing a decision, and the responses that follow
Decisions
What the analysis tells you
Every output exists to move a real decision (proceed, pause, adapt or stop), not to decorate a risk appendix.
Drivers
Which assumptions actually matter
The ranked critical drivers: the handful of assumptions that move the outcome, separated from the many that only look important.
Breakpoints
Where the thresholds sit
The specific values (of demand, price, cost, timing) at which the decision stops working, and how much headroom currently exists against each.
Range
The coherent outcome range
Upside, base, downside and stress outcomes built on internally consistent futures, so the board debates one honest range instead of duelling forecasts.
Failure
What failure would actually look like
Explicit failure conditions from the reverse stress test: the combinations that would sink the decision, and the early warnings that precede them.
Comparison
Options compared on a risk-adjusted basis
Where alternatives exist, they are compared after accounting for how badly each can go wrong, not just on their best-behaved base cases.
Actions
What management does, and when
The pre-agreed responses: which indicator, what trigger level, whose call, what action, so drift is met with a plan rather than a meeting.
Deliverables
What you receive
- ✓Critical assumption register: the drivers that matter, ranked, with plausible ranges
- ✓Scenario model with coherent upside, base, downside and stress cases
- ✓Stress-test results: what breaks, at what severity, in what order
- ✓Breakpoint analysis: the threshold values at which the decision stops working
- ✓Reverse stress findings: explicit failure conditions and early-warning indicators
- ✓Contingency playbook: indicator, trigger, owner, action, escalation, review frequency
- ✓Validation priorities: which assumptions deserve evidence-gathering before commitment
- ✓Governance recommendation: how and how often the framework is reviewed
How we work
How the work runs
The sequence is deliberate: failure gets defined before anything is stressed, and the model is judged fit for purpose before its outputs are trusted.
- 01
Decision and risk criteria
We agree precisely which decision is being tested, the approval timeline it must serve, and what failure means in its own terms: covenant breach, cash floor, payback ceiling, service level. Without this, stress testing is theatre.
- 02
Model review
The base model is examined for structural soundness: logic, data lineage and whether it can bear the weight the analysis will place on it.
GateIf the base model is unreliable, we say so and rebuild it first: stressing a broken model produces confident nonsense.
- 03
Sensitivity screening
Each assumption is flexed across its plausible range to rank the critical drivers and discard the noise, focusing everything that follows on what matters.
- 04
Scenario design
Coherent upside, base and downside futures are constructed with the sponsor and challenged for internal consistency, every variable moving the way it would in that world.
- 05
Stress and reverse stress
Severe-but-plausible shocks are applied and breakpoints measured; then failure is defined and traced backwards to the conditions and early warnings that would produce it.
- 06
Response planning
For each critical driver: indicator, trigger, owner, action, escalation and review frequency (the contingency playbook), agreed while everyone is calm.
- 07
Board review
Findings are presented in decision terms (proceed, pause, adapt or stop, and under what conditions) with the assumptions, thresholds and severity choices open to challenge.
Scope
What this is, and what it isn't
What's included
- All four analysis types (sensitivity, scenario, stress and reverse stress) applied as the decision requires
- An assumption and dependency map covering model logic, correlations and external drivers
- Failure defined explicitly in the decision's own criteria before testing begins
- Severity and plausibility of every stress stated and justified, never asserted
- A base-model quality review, with an honest rebuild recommendation if the model cannot support the analysis
- A contingency playbook with named owners, triggers and pre-agreed actions
- A board-ready summary framed as proceed, pause, adapt or stop
What this is not
- Stress testing on top of an unreliable base model: if the foundation is broken, we rebuild it first rather than stress a fiction
- Prediction: scenarios are coherent possibilities with explicit assumptions, not forecasts of what will happen
- Arbitrary worst-case pessimism disconnected from plausibility
- Formal audit or assurance of your financial statements or models
- Regulated financial product or investment advice
- A material investment, commitment or plan approval ahead
- Board or investor challenge to the base case that needs an analytical answer
- A forecast the leadership team is quietly split on
- Genuine volatility in costs, demand or funding conditions
- A recovery or turnaround plan that must survive its downside
- Willingness to define what failure means before testing begins
- A working model, or the appetite to build one properly first
Probably not the right fit
- −Anyone wanting analysis engineered to bless a decision already made
- −Expecting a prediction of the future rather than a tested range with thresholds
- −A base model known to be unreliable, with no appetite to fix it first
- −Organisations unwilling to act on triggers: a playbook nobody will execute is decoration
- −Decisions too small to justify the work: a rough sensitivity check may be all that's warranted
Adjacent services
The robustness layer around other decisions
Stress testing rarely stands alone. It hardens the models, forecasts and plans built elsewhere, and hands its findings to whichever pathway owns the response.
Sensitivity, Scenario & Stress Testingthe robustness layer
Owns: critical assumptions · coherent scenarios · failure thresholds · response triggers.
Sensitivity, Scenario & Stress TestingDemand Forecasting & Commercial Modelling
When it comes into play
The base forecast or model itself needs building (or rebuilding) before it can be stressed.
What it owns
Forecast construction, demand drivers and commercial model credibility.
Relationship to Sensitivity, Scenario & Stress Testing
Forecasting builds the base case; this work establishes where it breaks.
Proof
Selected work.
Payment Processing Cost Reduction. An ecommerce retailer was losing a significant percentage of revenue to payment processing and invoice platform fees. Web Lifter redesigned the entire sales and payment workflow, replacing Stripe and Paycove with a direct Westpac PayWay integration and a custom-built invoicing platform. The new architecture reduced transaction costs, streamlined operations, and delivered immediate profit improvements without requiring any increase in sales volume.
Read the caseWhat clients say
“We can't recommend Web Lifter highly enough … a digital partner who could understand our operations, connect the dots between marketing and backend systems, and deliver real results.”
Questions
Common questions
Method
Sensitivity analysis changes one assumption at a time to find which ones actually drive the outcome. Scenario analysis changes assumptions together, coherently, building whole futures where the variables move the way they would in reality. Stress testing pushes those futures to the severe end of plausible to see what breaks first. They answer different questions, which is why the work uses all three rather than picking one.
It runs the analysis backwards. Instead of asking 'what happens if things get bad?', it defines what failure means for this specific decision (a covenant breach, a cash floor, a missed payback) and works backwards to the combinations of conditions that would cause it. That surfaces failure paths forward-looking analysis misses, and identifies the early-warning indicators that would show up first.
The variables move together the way they do in the real world. A demand shock rarely arrives alone: it typically brings price pressure, slower customer payments and sharper competition with it. An incoherent downside ('revenue minus 20%, everything else unchanged') understates risk precisely because it ignores those correlations. Coherent scenarios are described futures, traced through every assumption they would touch.
Model & inputs
Then we tell you, and we rebuild it before stress testing anything. Stress results inherit every weakness of the model underneath them: stressing a broken model produces confident nonsense with a chart. The model review early in the process exists precisely to make this call honestly.
The decision or model being tested, the key assumptions as you currently understand them, the approval date the analysis must serve, any risk criteria that already exist, and whether a working model is available. That is enough to scope honestly, including telling you if the decision doesn't warrant the full treatment.
No, it's the first step of the process. Most organisations haven't written down what failure means for a given decision; defining it precisely, in the decision's own terms, is where the engagement starts and is valuable on its own.
Outputs & governance
Six things per critical driver: the indicator to watch, the trigger level at which action starts, the named owner, the pre-agreed action, the escalation path, and the review frequency. It converts analysis into governance: when reality drifts toward a threshold, management executes a plan instead of convening a debate.
The scope is set against your approval date: that's one of the first things we ask. The validation priorities also make the trade-off explicit: which assumptions are worth gathering evidence on before signing, and which can be governed by triggers and monitoring after. Robustness and momentum are sequenced, not traded off blindly.
A risk register lists things that could go wrong, usually with subjective likelihood and impact scores. This work quantifies where the decision actually breaks, ranks the assumptions that drive it there, and pre-agrees the management response at each threshold. The register names the weather; this tells you at what wind speed to take the sails down, and who does it.
Test the base case