skip to content

How does OWASP Risk Rating break likelihood and impact into factors you can grade?

level: middleimportance: nice to knowfreq 38%

answer

  1. Two groups of four, twice over
  2. Threat agent and vulnerability, then technical and business
  3. Skill, motive, opportunity, size
  4. Zero to nine, averaged per group
  5. Detection scores worst when nothing is reviewed

basics

~20 s

OWASP Risk Rating splits likelihood into four threat-agent factors and four vulnerability factors, and impact into four technical and four business factors. Each is scored 0 to 9 and averaged within its group, then read as low, medium or high.

solid answer

~50 s

Likelihood has two groups of four. Threat agent factors are skill level, motive, opportunity and size; vulnerability factors are ease of discovery, ease of exploit, awareness and intrusion detection. Impact also has two groups of four: technical impact is loss of confidentiality, integrity, availability and accountability; business impact is financial damage, reputation damage, non-compliance and privacy violation. Each factor is scored 0 to 9, averaged inside its group and read as low, medium or high, with the business figures preferred over the technical ones when you can get them. On a freight-broker load-tender portal where a rival broker can read other brokers' tender prices, motive and opportunity score high, skill moderate, size low because there are only a handful of competitors; discovery and exploit are easy; and if tender reads are logged but never reviewed, intrusion detection scores badly, which pushes likelihood up.

go deeper

for a junior

Be ready to say that likelihood and impact each split into two groups of four scored factors, and that the point is to make two raters argue about something specific rather than about a vague high or low.

for a middle

An interviewer expects the factor names and the scoring mechanics: 0 to 9, averaged within a group, read as low, medium or high. Know that intrusion detection is scored inversely and that it sits on the likelihood side.

for a senior

Show you can grade a real threat factor by factor and then explain which two or three factors drove the result. Demonstrate that you go and get the business impact numbers rather than substituting the technical group you can fill in alone.

for a principal

Own how the method is adapted: which factors your organisation drops or weights, how the option definitions are kept stable so ratings from different teams are comparable, and how you stop an averaged number from being read as a measurement in a steering meeting.

## Why a factor list at all The useful thing about OWASP Risk Rating is not the arithmetic, it is that it forces disagreement to become specific. Two people arguing about whether a threat is "high" get nowhere; two people arguing about whether *opportunity* is a 4 or a 9 are arguing about a checkable fact — what access the attacker must already have. The factors are a shared vocabulary for a conversation that otherwise runs on instinct. ## The four groups | Axis | Group | Factors | | --- | --- | --- | | Likelihood | Threat agent | skill level, motive, opportunity, size | | Likelihood | Vulnerability | ease of discovery, ease of exploit, awareness, intrusion detection | | Impact | Technical | loss of confidentiality, loss of integrity, loss of availability, loss of accountability | | Impact | Business | financial damage, reputation damage, non-compliance, privacy violation | Each factor is given a score from 0 to 9 using the option lists the method supplies, the scores in a group are averaged, and the average is read as **low** below 3, **medium** from 3 up to 6, and **high** from 6 to 9. Likelihood and impact are then read together to produce an overall severity for the threat. ## Grading a worked threat Take a freight-broker load-tender portal. The threat: a competing broker with a legitimate carrier account can enumerate tender records and read the prices and customer names another broker offered. Asset at stake is pricing and customer-list trade secrets. - *Skill level*: modest — an authenticated user changing an identifier, not an exploit developer. - *Motive*: high — the reward is directly commercial and repeatable every day. - *Opportunity*: high — the attacker already holds a valid account, so no special access is needed. - *Size*: low — the realistic attacker set is a handful of competing brokers, not the internet. - *Ease of discovery* and *ease of exploit*: high — sequential identifiers surface by accident. - *Awareness*: middling — the flaw is not published, but a competitor who finds it will not report it. - *Intrusion detection*: poor if tender reads are logged but nobody reviews them. On impact, the technical loss is confidentiality only — no integrity or availability effect. The business factors are where this threat is actually decided: financial damage from systematically undercut bids, reputation damage with shippers whose commercial terms leaked, and possibly contractual or non-compliance exposure. Note how business impact outranks the technical picture; a purely technical read would call "one field disclosed" minor. ## Two directions people get backwards **Intrusion detection is scored inversely.** A system with active detection and response scores at the low end of the range; one that logs without review scores mid-range; one that logs nothing scores at the top. A high score means the attacker is unlikely to be caught, which raises likelihood. Candidates frequently assume "good logging, so a high detection score", which reverses the whole factor. **Detection is a likelihood input, not an impact input.** It changes whether the attack proceeds and repeats, not what is lost when it succeeds. What logging protects on the impact side is a different factor entirely — loss of accountability, meaning whether actions can be traced to an individual. ## Business impact over technical impact The method is explicit that business impact is what matters and technical impact is a stand-in for it. Engineers reach for the technical group because they can fill it in alone; the business group needs a product owner, a commercial lead or legal to tell you what a leaked customer list or a missed regulatory duty actually costs. Doing that legwork is the difference between a rating an engineer believes and a rating that gets funded. ## Use it as a guide, not a measurement The method itself says the factors can be dropped, added or weighted for your context — a threat with no personal data has nothing to say about privacy violation, and forcing a number there just dilutes the average. Treat the resulting figure as a way to keep ratings comparable and the discussion anchored, not as a precise quantity. When you present the rating, present the two or three factors that drove it; the average alone tells the reader nothing about *why* the threat scored where it did, which is the only part that helps them decide what to do.

  • Why does the method tell you to prefer the business impact factors over the technical ones?
    Because technical impact is only a proxy. Loss of confidentiality does not tell anyone whether to fund a fix; financial damage, reputation damage, non-compliance and privacy violation do, and they are what the people approving the work actually weigh. Engineers default to the technical group because they can fill it in without leaving the room. Getting the business numbers means asking a product owner or a commercial lead, and that conversation is usually what makes the rating stick.
  • A system logs every request but nobody reads the logs. How does that score on intrusion detection?
    Badly, in the sense that the score goes up. The factor is scored inversely: active detection and response sits at the low end, logged-and-reviewed in the middle, logged-without-review higher, and not logged at all at the top. A high score means the attacker will not be noticed, which raises likelihood. The unread log still helps with loss of accountability on the impact side, because actions remain traceable after the fact.
  • What do you do when one of the factors does not apply to your threat?
    Drop it, and say you dropped it. The method explicitly allows removing, adding or weighting factors for your context. Forcing a zero into privacy violation for a threat that touches no personal data just drags the group average down and makes the rating less comparable, not more rigorous. The value of the list is the shared vocabulary and the record of what drove the score, not completeness of the arithmetic.

saying these in an interview costs you the question

  • Recites the factor names without applying any to the threat
  • Reads the averaged score as a probability
  • Puts financial damage into the likelihood group
  • Thinks a well-monitored system scores high on intrusion detection
  • Skips business impact because technical impact is easier to fill in
  • Treats the factor list as mandatory and unweightable

context