17 November 2026 • Cyber security
In brief: Security awareness metrics show whether training changes what people do, not just whether they clicked through a module. A useful set covers five layers: activity (participation, completion), behaviour (click rate, reporting rate and time to report in phishing simulations, repeat clickers), knowledge (test scores against a baseline), culture (survey results) and outcome (real incidents and reported suspicious emails). Of these, the reporting rate says the most, because it measures the behaviour that actually stops an attack. For the board, five to seven KPIs per quarter are enough, each with a trend and an action. The same figures serve as evidence for an ISO 27001 audit or a NIS2 supervisor.
Why participation alone proves nothing
Most security awareness programmes start by reporting one number: the percentage of employees who completed the training. It is easy to measure and it looks good on a slide. It also says nothing about risk. An employee can finish every module and still enter a password on a fake login page the next morning.
Auditors and regulators have noticed. ISO 27001 asks organisations to evaluate the effectiveness of the actions they take to build competence and awareness (clause 7.2) and to monitor and measure the performance of the management system (clause 9.1). The NIS2 Directive lists cybersecurity training among the minimum measures in Article 21(2)(g), and in Article 21(2)(f) it asks for policies and procedures to assess the effectiveness of those measures. A completion rate shows that training took place. It does not show that it works.
That is the gap a good security awareness program closes: from “everyone did the course” to “here is what changed, and here is what we do next”.
Five layers of security awareness metrics
It helps to sort metrics by what they actually tell you. The higher the layer, the closer the number gets to real risk, and the harder it is to measure.
1. Activity: did people take part?
Participation rate (who started) and completion rate (who finished, on time). These are hygiene metrics. You need them, because training that nobody follows changes nothing, and because NIS2 and DORA expect you to show that staff and management were trained. But they measure effort, not effect.
2. Behaviour: what do people do when it counts?
This is where phishing simulations come in. They are not mandatory under any law, but they are the most practical way to see behaviour instead of intentions. Four numbers matter:
- Click rate: the share of recipients who clicked the link or opened the attachment.
- Data entry rate: the share who went further and entered a password or other data.
- Reporting rate: the share who reported the email to the security team or through a report button.
- Time to report: how quickly the first reports come in after the email lands.
Add a fifth: repeat clickers, the people who fail several simulations in a row. They are a small group in most organisations, and they deserve extra support rather than a public list.
3. Knowledge: do people understand the threats?
Short knowledge tests on themes such as phishing, passwords, CEO fraud and data handling. On its own a test score is a snapshot. Compared with a baseline measurement taken before the programme started, it becomes a trend you can report.
4. Culture: do people dare to speak up?
A short survey with questions such as “Would you report it if you clicked a suspicious link?” and “Do you know where to report?”. Culture is the layer that explains the others: if people fear blame, the reporting rate stays low however good the training is. Our article on building a security awareness culture goes deeper.
5. Outcome: what happens in the real world?
The number of real suspicious emails reported by employees, the share of real incidents with a human cause, and the time between a real phishing email arriving and the security team knowing about it. These numbers are noisier and depend on more than training, but they are what the board ultimately cares about.
Why the reporting rate matters more than the click rate
A click rate of zero does not exist. Phishing emails are getting better, and in a large organisation someone will always click. What decides the damage is how fast the security team hears about it. One early report lets IT block the sender, pull the email from every mailbox and reset the password of the colleague who clicked, before the attacker gets further.
The numbers point the same way. According to the Verizon Data Breach Investigations Report, the human element plays a role in around 60% of breaches (2025 edition). The 2025 edition also found that employees with recent training reported simulated phishing at a rate of about 21%, against 5% for employees without recent training, while the effect on the click rate was much smaller. In other words: training moves reporting faster than it moves clicking.
There is a second reason. The click rate is easy to manipulate (more on that below), and it rewards people for doing nothing. The reporting rate rewards an action. An employee who clicks and then reports straight away has helped the organisation. Good security awareness reporting shows that difference.

The KPI table: what to measure and for whom
The table below sets out the most useful KPIs. Public benchmarks are scarce and the figures that circulate mostly come from vendors with their own customer base and difficulty levels. That is why the target value for most KPIs is your own baseline: what counts is the trend inside your organisation.
| KPI | What it measures | How to calculate it | Target or benchmark | For whom |
|---|---|---|---|---|
| Participation rate | Reach of the programme | Employees who started ÷ employees in scope × 100 | Own target, for example (close to) everyone in scope; management body included under NIS2 | Board, HR |
| Completion rate | Whether training is finished on time | Assigned training completed within the period ÷ assigned training × 100 | Own target; trend per department | HR, compliance |
| Click rate | Susceptibility to phishing | Recipients who clicked ÷ recipients who received the simulation × 100 | Lower than own baseline at comparable difficulty; no reliable public benchmark | IT, security |
| Data entry rate | How often a click leads to real damage | Recipients who entered data ÷ recipients × 100 | Falling against own baseline | IT, security, board |
| Reporting rate | Whether people raise the alarm | Recipients who reported ÷ recipients × 100 | Rising against own baseline; Verizon DBIR: about 21% with recent training, 5% without (2025) | Board, IT, security |
| Time to report | How fast the security team gets a warning | Median minutes between delivery and first report (or first ten reports) | Shorter than own baseline; benchmark: the time it takes the first person to click, which according to Verizon is usually less than a minute (DBIR 2024) | IT, security |
| Repeat clickers | Concentration of risk | Employees who failed two or more of the last X simulations ÷ employees in scope × 100 | Falling; follow up with extra training, not blame | IT, HR |
| Knowledge score | Understanding of threats per theme | Average test score per theme, compared with the baseline measurement | Higher than baseline; lowest themes are the next priority | HR, IT |
| Culture score | Willingness to report and ask questions | Average agreement on fixed survey questions, once or twice a year | Rising against previous survey | Board, HR |
| Real reports of suspicious email | Behaviour outside simulations | Number of real reports per month, and the share that turned out to be malicious | Rising number, stable or rising share of real threats | IT, security, board |
| Incidents with a human cause | Outcome in practice | Security incidents with phishing, social engineering or human error as cause, per quarter | Falling, or detected earlier | Board, compliance |
Guardey trains employees in 3-minute weekly challenges and reports participation and progress per employee, with scores per topic and weekly reports.
Start with a baseline
A trend needs a starting point. A baseline measurement combines an unannounced phishing simulation with a short knowledge test and, optionally, a culture survey. You analyse the results per team, department or location rather than per person, so the measurement stays about the organisation and not about who to blame. Then measure again every quarter with the same metrics and comparable difficulty. After a year you have five data points, and that is a story a board can follow.
How to build a quarterly report for the board
Directors do not need eleven KPIs. They need to know whether the risk is going down, where it is not, and what is being done about it. A one-page quarterly report works best:
- Pick five to seven KPIs and keep them fixed. For example: participation rate, reporting rate, data entry rate, time to report, knowledge score, real reports of suspicious email and incidents with a human cause. Changing the set every quarter makes trends impossible.
- Show the trend, not the snapshot. Each KPI gets the current value, the previous quarter and the baseline. An arrow says more than a percentage.
- Explain the outliers. One department with a high data entry rate, or a theme where the knowledge score lags. Two sentences each.
- Attach an action to every red number. Extra training for one team, a simulation on a theme that scored low, a reminder about where to report. Without an action a report is only a scoreboard.
- Record board participation separately. Under NIS2 Article 20(2), members of the management body must follow training themselves. Report their participation as its own line, and consider NIS2 training for executives if that line stays empty.
Keep a written security awareness training policy next to the report: who is in scope, how often people train, how simulations are run and how results are used. That policy explains the numbers to anyone who reads them later.
Pitfalls: when metrics look better than reality
Vanity metrics
Hours of training, number of modules launched, number of posters on the wall. They show activity and they always go up. A board that only sees these numbers concludes that the risk is under control, while nobody knows whether behaviour has changed.
Making the click rate look good with easy simulations
The quickest way to halve your click rate is to send obvious phishing: spelling mistakes, a strange sender, a prince who wants to share his fortune. The figure improves, the risk does not. Keep the difficulty comparable between measurements, record the difficulty level with every result, and vary the scenarios so people do not learn to recognise one template.
Measuring individuals instead of the organisation
Naming and shaming repeat clickers feels decisive, but it pushes people to stay quiet after a mistake, and then the reporting rate drops. Use individual data to offer extra support, and report to the board at team or department level.
Counting without context
A rise in real reports of suspicious email can mean more attacks or more alert employees. Look at the share that turned out to be malicious and at what happened in the threat landscape that quarter before drawing a conclusion.
Metrics as evidence for auditors and supervisors
An auditor or supervisor does not ask for perfect scores. They ask whether you know where you stand, whether you act on it and whether you can show that over time. That is exactly what a fixed set of metrics with a baseline and quarterly trend gives you.
- ISO 27001: Annex A control 6.3 requires security awareness, education and training for personnel, and clauses 7.2, 7.3 and 9.1 ask you to evaluate whether that training works. Participation records, knowledge scores against a baseline and simulation trends are the evidence. Our page on ISO 27001 security awareness training shows how that fits together.
- NIS2: training for staff (Article 21(2)(g)) and for the management body (Article 20(2)), plus an assessment of whether the measures are effective (Article 21(2)(f)). The quarterly board report with actions is the paper trail.
- DORA: financial entities must run ICT security awareness programmes and digital operational resilience training for all staff and the management body (Article 13(6)). The same KPIs show that the programme exists and is evaluated.
Keep the evidence together: the training policy, the baseline, the quarterly reports and the actions you took. When an auditor arrives, that file answers most questions before they are asked.
Frequently asked questions about security awareness metrics
What are the most important security awareness metrics?
For most organisations: participation rate, reporting rate, data entry rate and time to report in phishing simulations, knowledge score against a baseline, and the number of real suspicious emails reported. Together they show reach, behaviour, knowledge and the effect in practice.
What is a good phishing click rate?
There is no reliable public benchmark, because difficulty and scenarios differ per organisation and per vendor. Compare your click rate with your own baseline at comparable difficulty, and pay more attention to the reporting rate and the data entry rate.
How do you measure the effectiveness of security awareness training?
Start with a baseline measurement: a phishing simulation, a short knowledge test and optionally a culture survey. Repeat it every quarter with the same metrics and comparable difficulty, and report the trend per department together with the actions you take.
Are phishing simulations mandatory under NIS2?
No. NIS2 requires cybersecurity training and an assessment of whether measures are effective, but it does not prescribe phishing simulations. They are a practical way to measure behaviour, not a legal obligation.
In a personal demo we walk through the reports per employee and department, based on the KPIs your board asks for.