Explore our new Cybersecurity Terms Glossary, explained by our experts.

View the Glossary View the Glossary
PRIVIA

Cybersecurity Terms Glossary

What Is the DREAD Risk Scoring Model?

DREAD rates a threat on damage, reproducibility, exploitability, affected users, and discoverability, turning a list of findings into a ranked queue of work.

Illustration for the cybersecurity term DREAD
Home / Blog / Cybersecurity Terms / What Is the DREAD Risk Scoring Model?
17 November 2026

What Is the DREAD Risk Scoring Model?

What Is the DREAD Risk Scoring Model?

DREAD is a five-factor scheme for ranking threats, built at Microsoft in the early 2000s as the companion to STRIDE. STRIDE tells you what could go wrong; DREAD tells you which of those things to fix first. The acronym stands for damage, reproducibility, exploitability, affected users, and discoverability.

The Five Factors

FactorQuestion it answersLow scoreHigh score
DamageHow bad is it if this happens?A non-sensitive detail leaksFull database disclosure or domain compromise
ReproducibilityHow reliably does the attack work?Needs a rare race conditionWorks every time with one request
ExploitabilityWhat does the attacker need?Insider access and custom toolingA browser and no account
Affected usersHow many people are hit?One administratorEvery customer of the service
DiscoverabilityHow easily is the flaw found?Requires source code reviewVisible in the page response

Each factor is scored on a small scale, most often 1 to 3 or 1 to 10, and the five values are added. On the 1–10 version a threat scores between 5 and 50; teams usually treat 40 and above as critical, the twenties and thirties as medium, and single figures as informational. The arithmetic is deliberately trivial—the value is in forcing a conversation about each factor rather than arguing about a single gut-feel label.

A Worked Example

Take an API endpoint that returns another customer’s invoice when the identifier in the URL is changed, a classic case of broken access control. Damage is high because invoices carry names, addresses, and amounts. Reproducibility is maximal: change a number, get a document. Exploitability is high because any authenticated user can do it. Affected users covers the whole customer base. Discoverability is high, since the identifier is a visible integer. On a 1–10 scale that lands near 45 and goes to the top of the queue, above a stored cross-site scripting flaw in an internal admin page that only three staff can reach.

Why Microsoft Retired It

Microsoft stopped recommending DREAD for its own threat modelling around 2008. The problem was consistency: two analysts reviewing the same flaw routinely produced scores ten points apart, because nothing in the model defines what a 6 means for damage. Discoverability was the worst offender—it rewards obscurity, so a team could lower a critical finding’s score by arguing that nobody would notice the flaw, which is security through obscurity with a number attached.

Averaging rather than summing created a second problem. A threat with catastrophic damage but low discoverability can average into the middle of the list, next to something trivial, and disappear.

DREAD and CVSS

For vulnerabilities in shipped software, CVSS has largely replaced DREAD. CVSS v4.0, released in November 2023, defines every metric value precisely, is maintained by FIRST, and produces a score that means the same thing to a supplier, a customer, and a regulator. DREAD has no such shared dictionary.

DREAD still has a niche. It is quick, it needs no tooling, and it works on things CVSS was never designed for: a business process, an abuse case, a design decision that has not been built yet. During a design workshop, scoring six candidate threats on five factors takes minutes and produces a defensible order of work.

Making It Usable

Whichever scheme you choose, the output has to feed a real risk assessment and a remediation plan, not a spreadsheet nobody reopens. Our penetration test services deliver findings with a documented scoring rationale, so your team can argue with the reasoning rather than guess at the number.

FAQ – Frequently Asked Questions

Is DREAD Still Used?

Rarely as a formal standard, but often as a quick workshop technique. Microsoft dropped it from its own guidance because scores varied too much between analysts. Teams that still use it tend to write a local rubric, remove or invert discoverability, and reserve it for design-stage threats and abuse cases where CVSS does not apply.

What Is the Difference Between DREAD and CVSS?

CVSS is a maintained open standard with precisely defined metric values, aimed at vulnerabilities in released software, and its scores are comparable across organisations. DREAD is an informal five-factor heuristic with no shared definitions, aimed at ranking threats during design. CVSS is better for patch prioritisation, DREAD is faster for threat modelling workshops.

Why Is Discoverability Controversial?

Because it lowers the score of a flaw simply because it is hard to notice, which is security through obscurity. An attacker who does find it is unaffected by how long the search took. Most practitioners either remove the factor, score it out of 40 instead of 50, or invert it so hidden flaws are not treated as safer.

How Does DREAD Relate to STRIDE?

They were designed as a pair. STRIDE enumerates threats by category—spoofing, tampering, repudiation, information disclosure, denial of service, and elevation of privilege—while DREAD ranks the threats that enumeration produces. Many teams now keep STRIDE for identification and swap DREAD for CVSS or a simple likelihood-and-impact matrix for ranking.

Related Terms

You May Be Interested In These

What Is a Lessons Learned Review?

Cybersecurity Terms

17 November 2026

What Is a Lessons Learned Review?

Read More
What Is Data Loss Prevention (DLP)?

Cybersecurity Terms

17 November 2026

What Is Data Loss Prevention (DLP)?

Read More
What Are Adversarial Examples?

Cybersecurity Terms

17 November 2026

What Are Adversarial Examples?

Read More