MTTD (mean time to detect) is the average time between the start of a security incident and the moment your team detects it. MTTR (mean time to resolve) is the average time from detection until the incident is fully resolved. MTTD tells you how good your visibility is. MTTR tells you how good your response is.
The two are often reported together, and they fail for different reasons. A team with excellent responders and poor logging will post a fine MTTR and a terrible MTTD, because nobody can fix an incident they have not noticed yet.
MTTD vs MTTR at a glance
| MTTD | MTTR | |
|---|---|---|
| Stands for | Mean time to detect | Mean time to resolve (see the note on the R below) |
| Clock starts | When the incident begins | When the incident is detected |
| Clock stops | When the team detects it | When the incident is fully resolved |
| What it measures | Visibility: logging, detection rules, monitoring coverage | Response: triage, diagnosis, repair, recovery, coordination |
| A rising number suggests | Gaps in logging, detection rules, staffing, or visibility | Delays in diagnosis, repair, coordination, or recovery |
| Improved by | Better log coverage, tuned alerting, threat hunting | Playbooks, clear ownership, practiced escalation, tested restores |
How to calculate MTTD and MTTR
Both are simple averages over a defined period, such as a month or a quarter.
- MTTD = total time between incident start and detection, across all incidents in the period, divided by the number of incidents.
- MTTR = total time between detection and resolution, across all incidents in the period, divided by the number of incidents.
A worked example. In one quarter a team handles three incidents. Detection took 2 hours, 10 hours, and 30 hours, so MTTD is (2 + 10 + 30) / 3 = 14 hours. Resolution after detection took 4 hours, 8 hours, and 24 hours, so MTTR is (4 + 8 + 24) / 3 = 12 hours.
The hard part of MTTD is the start time. You rarely know when an incident began until the investigation is over, so the start timestamp usually comes from forensics: the first malicious log entry, the first phishing click, the first unauthorized login. That is why MTTD depends on log retention. If the logs from the start of the intrusion have already been deleted, you cannot measure it.
The R in MTTR has four meanings
MTTR is used for four different measurements, and reports rarely say which one they mean. Atlassian’s incident management guide separates them clearly:
| Term | What it covers |
|---|---|
| Mean time to repair | The time spent actively repairing and testing a fix |
| Mean time to recovery | The time from the start of an outage until service is fully restored |
| Mean time to respond | The time from the first alert until recovery, excluding any lag before the alert |
| Mean time to resolve | The time to fully resolve the incident, including work done to stop it happening again |
In security work, mean time to resolve is usually the useful one, because an incident is not over when the service comes back. It is over when the root cause is fixed. Whichever you pick, write the definition next to the number and keep it the same from one report to the next.
Common mistakes
- Reporting one blended number. A single MTTR across every incident hides the difference between a phishing email and a ransomware event. Split the metric by severity, and where it helps by business unit or detection source.
- Treating recovery as one event. After a large disruption, different business services come back at different times. Measuring MTTR for each service shows which recovery processes worked and which did not.
- Letting one outlier set the average. With few incidents, one long investigation dominates a mean. Report the median next to it.
- Counting only what was detected. MTTD can only include incidents you found. A low MTTD alongside low log coverage may mean you are missing incidents, not catching them quickly. Pair MTTD with a coverage metric such as audit log coverage.
- Chasing the number. Closing tickets early improves MTTR on paper and makes security worse. Define “resolved” so that it includes verifying the fix.
What these metrics are for
Neither number is useful as a one-off. Their value is the trend. A falling MTTD after you deploy new detection rules is evidence that the investment worked. A rising MTTD is an early warning that logging, staffing, or visibility has slipped while attackers get more time inside.
NIST’s incident response guidance, SP 800-61, treats detection, response, and recovery as parts of one continuous risk management cycle. MTTD and MTTR are how you check whether that cycle is speeding up.
See it applied
Frequently asked questions
- What is the difference between MTTD and MTTR?
- MTTD is the average time from the start of an incident until it is detected. MTTR is the average time from detection until the incident is resolved. MTTD measures visibility, and MTTR measures response.
- How do you calculate MTTD?
- Add up the time between incident start and detection for every incident in a period, then divide by the number of incidents. The start time usually comes from forensic evidence such as the first malicious log entry.
- What does the R in MTTR stand for?
- It can mean repair, recovery, respond, or resolve, and each one measures a different span of time. In security reporting, mean time to resolve is the most common. Always state which definition you are using.
- What is a good MTTD?
- There is no universal target, because it depends on the type of incident and the environment. The useful comparison is your own trend: whether MTTD is falling over time for incidents of the same severity.