Explore our new Cybersecurity Terms Glossary, explained by our experts.
Cybersecurity Terms Glossary
Return on security investment, or ROSI, estimates the value of a control by comparing the loss it is expected to avoid with what it costs to buy, run, and operate.

What Is Return on Security Investment?
Return on security investment is the attempt to express a control in the language finance uses for everything else: what does it cost, and what does it save. The standard formula compares the loss a control is expected to prevent with its total cost of ownership, and the result is quoted as a percentage.
Security leaders reach for ROSI when a board asks why a detection platform costs more than the incident it was bought to prevent. Used carefully it structures that conversation. Used carelessly it produces a number with four significant figures and no credibility.
ROSI builds on annualised loss expectancy. Single loss expectancy (SLE) is the cost of one occurrence; annual rate of occurrence (ARO) is how often you expect it; multiply them for the annualised loss expectancy (ALE). The control is then credited with removing some share of that loss:
ALE = SLE x ARO
ROSI = ((ALE x mitigation ratio) - cost of control) / cost of control
Take a business email compromise scenario: an expected loss of £180,000 per successful incident, an estimated 0.5 occurrences a year, giving an ALE of £90,000. A combination of strong authentication and payment verification is judged to stop 80% of it, saving £72,000 a year against a £25,000 annual cost. ROSI is therefore (72,000 − 25,000) / 25,000, or 188%.
Every input is an estimate, and some are better than others. Cost of control is the solid one: licences, implementation, and the staff time to run it, which finance can verify. The loss side is harder, and the honest approach is to show your sources:
The mitigation ratio is usually the weakest link. Claiming that a tool stops 95% of an attack class is a vendor number; evidence from your own testing, including tabletop exercises and purple team results, is worth more than any datasheet.
Economists Lawrence Gordon and Martin Loeb showed in 2002 that the optimal investment in protecting an information set is generally well below its expected loss, with the widely quoted ceiling at roughly 37%. The intuition is simple: the first controls remove the easy attack paths cheaply, while the last few percent of residual risk costs disproportionately more to remove. ROSI should fall as you spend more on the same risk, and a model that shows ever-rising returns is wrong.
Loss avoidance is only part of the case. Several of the strongest arguments for security spending never appear in the formula:
| Benefit | How to evidence it |
|---|---|
| Revenue enablement | Deals won or unblocked because you hold ISO 27001 or passed a client’s security review |
| Regulatory exposure avoided | UK GDPR fines up to £17.5 million or 4% of global turnover, plus the cost of ICO engagement |
| Insurance terms | Premium or excess changes when MFA and tested backups are in place |
| Operational efficiency | Engineer hours saved by automation, measurable from ticket data |
These are easier to defend than probability estimates, because the evidence already exists in your sales pipeline and renewal paperwork.
Give boards a range, not a point. State the assumptions on the same slide as the result, show how the answer changes if the frequency estimate halves, and tie the scenarios to risks already on the register so the conversation connects to your risk appetite rather than running parallel to it. Used this way, ROSI becomes a tool for comparing options against each other, which is what budgeting actually requires, rather than a claim of absolute return.
If your loss estimates need a stronger evidence base, our cyber risk analysis services quantify exposure using real technical findings rather than generic averages.
It is, provided the assumptions are visible. Finance teams are used to appraising investments with uncertain returns and will accept ranges and sensitivity analysis. What damages credibility is a single precise percentage with no stated basis, or a model that only ever recommends buying something. Showing where a control stops paying for itself tends to increase trust rather than reduce budget.
Annualised loss expectancy is the expected annual cost of a risk before any new control, calculated as single loss expectancy multiplied by annual rate of occurrence. ROSI goes a step further by estimating how much of that loss a specific control removes and comparing the saving with the control’s total cost. ALE describes the problem; ROSI appraises a proposed answer to it.
Build it from components rather than copying an industry average: incident response and forensics, legal advice, customer notification, regulatory engagement, downtime measured against revenue per hour, and remediation. Add churn only where you can evidence it. Your own past incidents, including near misses and the hours they consumed, are usually a better guide than global figures weighted by very large breaches.
Yes. Controls required by law, by a regulator, or by a major customer contract are not optional, whatever the arithmetic says. The same applies to controls that enable revenue, such as the certification a client demands before signing. ROSI is one input to a prioritisation discussion, not a veto over compliance obligations or commercial commitments.
You May Be Interested In These
We use a few cookies that are needed to make this website work, and they are always on. We would also like to set analytics cookies, to see which pages are useful; functional cookies, so that our live chat works; and advertising cookies, to measure our campaigns. We will not set any of those until you say yes, and you can change your mind at any time from the Cookie Settings link in the footer.
Choose the cookies you are happy for us to use. Strictly necessary cookies are always on; everything else is off until you turn it on.
These cookies keep the website working: they hold your security checks, your form submissions and the choice you make here. The site cannot run without them, so they are always on.
These let our live chat open and remember your conversation, and let the newsletter sign-up work. Switch them off and the chat window will not appear.
These tell us which pages are read and how people find us, so that we can improve the site. The reports are about groups of visitors, never about you personally.
These let us measure our advertising and show you our adverts on other websites, such as LinkedIn, Google and Meta. Switch them off and you will still see adverts, but they will be less relevant.