INSIGHTS / APPLIED USE CASES

USE CASE · RISK ANALYSIS

The Hardest Part of Risk Analysis Comes Before the Math

Your CISO asks you to quantify ransomware risk. It sounds simple, until you sit down to do it.

Ransomware on what? The payments platform, or a file server nobody has touched since 2019? Launched by whom, through what door? Does the loss come from four days of downtime, from paying the ransom, or from the regulator who shows up afterward? Pick one answer to each question and you have a scenario. Pick different answers and you have a different scenario, with a different cost.

Most conversations about risk quantification go straight to the data, the ranges and the simulation. Those problems are real. They also assume you already know what you're measuring, and that turns out to be harder to settle than it sounds.

A wide field of scattered points with a small cluster highlighted, illustrating a few scenarios chosen out of thousands of possible ones.

Two Analysts, Two Numbers

Hand the same request to two good analysts. One models an attacker encrypting the payments platform. The other models a stolen laptop with customer records on it. Both did careful work. Both labeled it "cyber risk." Their numbers are hard to compare, hard to add together, and hard to use when you're deciding what to fund. Neither made a math mistake. Each answered a reasonable question, just a different one, because the request left room for both.

One request, four scenariosA single request to quantify ransomware risk fans out into four different scenarios: the payments platform encrypted and down for days, costing lost revenue and recovery; an old file server encrypted with nothing urgent breaking, costing cleanup; customer records stolen before encryption, costing notification and regulator response; and the payments platform with a ransom paid, costing payment and legal. Each carries a different cost, producing four numbers that are hard to compare.The request: quantify ransomware riskPayments platformEncrypted, down for daysCost: lost revenue and recoveryDIFFERENT COSTOld file serverEncrypted, nothing urgent breaksCost: cleanupDIFFERENT COSTCustomer recordsStolen before encryptionCost: notification and regulatorresponseDIFFERENT COSTPayments platformRansom paidCost: payment and legalDIFFERENT COSTFour scenarios, four numbers that are hardto compare.
One request can mean four very different scenarios, and each carries its own cost.Illustrative

Choosing Which Scenarios to Model Is Hard

Thousands of things could go wrong at your company, and every team has to choose which ones to model. Choices often come down to whoever has the strongest view, whatever caused trouble last quarter, or whatever data happens to be on hand. That last one is easy to fall into. You end up modeling what's easy to measure, and some of what could hurt you most goes unmodeled.

Then you have a register of forty scenarios. Do any two describe the same loss? Is anything missing? It's hard to say, because there's rarely a clear picture of what "complete" would look like.

Start From the Decision

Say you leave a known vulnerability open on one server for 90 days. The scenarios worth modeling come from that decision: what the open vulnerability makes possible, on that server, while the acceptance stands. That question is small enough to answer.

Meridian starts from the decision. The scenarios worth modeling come out of what changed in the business, with no blank page to fill.

APPLIED USE CASES