Risk analysis software quantifies uncertainty in a decision, a budget or a schedule. Rather than producing a single figure, it represents uncertain inputs as ranges and probabilities, combines them — usually by Monte Carlo simulation — and reports the distribution of possible outcomes, so a plan can be funded to a stated confidence level.
The practical difference is what you can say afterwards. Without it, a cost estimate is a number and an argument. With it, the estimate is a number attached to a probability: this budget holds in four cases out of five. That is a claim a programme board can interrogate, and a claim you can be held to honestly, which is why quantitative risk analysis is mandated in sectors from oil and gas to public infrastructure.
The difference it makes is easiest to see. A single-figure estimate is one point on a distribution the model can draw in full — and the gap between that point and the confidence level you actually want to fund is the contingency:
It is worth being precise about the boundary. Software that records risks — a register of what might go wrong, who owns it and how likely someone thinks it is — is a different category from software that models them. Both are useful. Only the second one tells you how much contingency to hold.
Four, and they answer different questions. Most of the frustration in this market comes from buying one type expecting it to do another's job. The quickest way to place your own problem is to ask what you are actually modelling:
In more detail, and with the people who typically buy each one:
A register and a simulation tool are complements, not alternatives: the register is where risks are captured and owned, the model is where their combined effect is quantified. The strongest setups feed one into the other — a quantitative risk register that carries probability and impact distributions can be simulated directly rather than summarised into a heat map.
Most organisations genuinely need two of the four. A construction contractor typically needs cost simulation and schedule risk. An investment team typically needs cost simulation and decision analysis. Almost nobody needs a single product that claims to do all four, and products that claim it usually do one of them well.
Feature lists are long and mostly interchangeable. These six capabilities are the ones that separate a model you can defend from one that merely produces a chart:
Fitting is the one most often skipped where data exists. ModelRisk ranks candidate distributions against the data on goodness-of-fit statistics rather than leaving the shape to judgement:
Auditability deserves particular weight, because it is where the difference between tools shows up months later. If the model lives in a proprietary file that only the licence holder can open, every review becomes a meeting. If it lives in an ordinary spreadsheet, a colleague, an auditor or a client can open it and follow the logic cell by cell.
A spreadsheet on its own gives you one scenario at a time. If the question is which single outcome to plan for, that is enough. If the question is how likely the plan is to hold, you need simulation — and then the choice is between dedicated software and building the sampling by hand.
Hand-built simulation is possible and occasionally the right answer for a one-off. It is also slow, hard for anyone else to audit, and easy to get subtly wrong in ways that do not announce themselves: correlation omitted, the same random draw reused where it should not be, a distribution truncated by accident, too few iterations for the tail you are quoting. The failure mode is not an error message. It is a confident number that happens to be wrong.
The middle path — and the reason spreadsheet add-ins are the most widely used category — is software that adds simulation to the spreadsheet you already have, leaving the model in Excel where the data, the reviewers and the audit trail already live.
Published list prices for named-user annual licences generally run from a few hundred to a few thousand euros per user per year. Network, enterprise and academic licensing is usually quoted rather than listed, and a significant number of vendors publish no price at all — which makes like-for-like comparison difficult before you are already in a sales process.
Three things move the real total well away from the headline figure. Edition ladders: where distribution fitting, correlation or optimisation sit in a higher tier, the edition you actually need is rarely the one quoted. Seat count: per-user pricing that looks reasonable for one analyst compounds across a team. Renewal: a first-year discount that does not persist changes a five-year comparison completely.
We publish our prices for exactly this reason. ModelRisk is €1 550 per user per year at most, in one edition with everything included; Tamara is €2 150; ModelChoice is €950. For a like-for-like comparison against other tools, with sourced prices and retrieval dates, see our reviews of the top Excel risk analysis add-ins and the top project risk analysis tools.
Start from the decision you need to defend, not from the feature matrix. Four questions settle most evaluations:
One question worth adding if the evaluation is genuinely balanced: it is possible to calculate what further investigation is worth before committing to it. That is value-of-information analysis, and it applies to buying decisions as readily as to engineering ones.
Vose Software has built quantitative risk tools since 2001, one product per category rather than one product claiming every category:
Schedule risk in Tamara, a rolled-back decision tree in ModelChoice, and a quantitative heat map in Pelican — where each plotted risk carries a simulated financial exposure rather than a colour band:
All of them keep the model somewhere a reviewer can reach it, and all are available on a fully functional 15-day trial. If you are not sure which category your problem falls into, the four questions above will place it in a few minutes.
Software that quantifies uncertainty in a decision, budget or schedule. It represents uncertain inputs as ranges and probabilities, combines them by simulation, and reports the distribution of outcomes — so a plan can be funded to a stated confidence level rather than to a single guess.
Four: spreadsheet simulation add-ins, schedule risk analysis tools, decision analysis tools, and risk registers. Most organisations need two of the four. Products claiming all four usually do one of them well.
Published named-user list prices generally run from a few hundred to a few thousand euros per user per year, with network and enterprise licensing quoted rather than listed. Many vendors publish no price at all, which makes comparison difficult before entering a sales process.
A spreadsheet gives you one scenario at a time. If you need to know how likely the plan is to hold, you need simulation — either dedicated software or hand-built sampling, which is slow, hard to audit and easy to get subtly wrong.
Represent uncertainty with appropriate distributions, fit distributions to data, correlate inputs that move together, model discrete risk events separately from ranges, simulate to a stable answer, and report percentiles and sensitivity in an auditable form.
Build the same small model in each on a trial licence, ideally one whose answer you already know. Compare how the model is expressed, whether a colleague could open and audit it, how fast ten thousand iterations run, and the total cost at the seat count you actually need.
Distribution fitting, correlation, risk events and full simulation results, in the Excel workbook you already have. One edition, everything included, published price.