How to Measure and Record Uptime and MTTR KPIs
Facilities teams often have two credible-looking spreadsheets and still cannot explain a change. One uses ticket closure dates; the other uses device events. The gap is usually a definition problem rather than an arithmetic problem. For uptime and mttr kpis measurement, start by fixing the population, event boundaries, and evidence hierarchy. Then select a source architecture.
Direct answer. The short answer is that uptime and mttr kpis measurement is not one decision. It is a sequence. This article takes uptime and mttr kpis measurement from uptime and mttr kpis measurement through uptime-and-mttr-kpis to the record you keep. An uptime and mttr kpis measurement is the operating-life view. It starts after the invoice. uptime and mttr kpis measurement is the first interval of uptime and mttr kpis measurement to measure.
uptime and mttr kpis measurement is taken here as uptime and mttr kpis measurement, deliberately narrowly. It is a working definition of uptime and mttr kpis measurement. It is specific to this market and product, and it is not a statutory or certified term.
- Choose the schedule — define how often uptime and mttr kpis measurement is inspected
- Choose the cutoff — name the condition that forces a change to uptime-and-mttr-kpis
- Report the reading — keep one metric that shows four measurement approaches enter is healthy
- For uptime and mttr kpis measurement, uptime and mttr kpis is what each step has to answer to.
This article compares four practical choices: manual field logs, service-ticket records, device telemetry, and a reconciled hybrid. It also defines the required calculation rules, minimum record, audit tests, and decision criteria for post-award portfolio reporting. No method is universally superior. Manual records can be direct but incomplete, tickets can provide governance but arrive late, and telemetry can add context while inheriting synchronization and inference gaps.
Four measurement approaches enter the ring
The comparison must begin with the event that starts the clock, not with the software used to display the result. Each approach creates a different audit trail between a reported fault and a door that is operable again.
| Measurement approach | Primary event source | Strongest feature | Main weakness | Best operating fit |
|---|---|---|---|---|
| Manual field service KPI log | Technician-entered timestamps | Direct evidence at the door | Missing or late entries | A controlled estate with a disciplined service routine |
| Service-ticket record | Ticket creation and update fields | Workflow accountability and searchable history | Closure can be mistaken for restoration | Teams already operate a formal service desk |
| Device telemetry | Device-generated state events | Frequent state observations without a site visit | Clock, connection, and state-inference gaps | Estates with a usable, governed event feed |
| Reconciled hybrid | Ticket spine supported by door verification | Traceable chain from report to restoration | More administration and reconciliation | Portfolio reporting across sites and contractors |
A hybrid approach is not permission to average conflicting sources. It requires a written rule assigning each event to an authoritative source. For example, the service desk may own the report time while the technician owns the operability verification. A method becomes defensible when another person can reproduce the same KPI from the retained evidence, not when the dashboard produces a stable-looking percentage. The product family this applies to is the smart door locks.
Draw the metric boundary before collecting the data
Use consistent definitions across every site and contractor. The following boundaries determine which clock is being read and prevent availability, response, attendance, and restoration figures from being presented as substitutes for one another.
| Measure | Starts | Stops | Shows | Does not show |
|---|---|---|---|---|
| Response time | Fault reported through the agreed channel | A person or system acknowledges it | Whether the report entered the response process | Whether anyone attended or repaired the door |
| On-site attendance time | Acknowledgement | Someone is at the door | Progress between acknowledgement and attendance | Whether the fault was fixed |
| Resolution time | Fault reported | Door is operable again | The end-to-end restoration interval | Why the restoration took that interval |
| MTTR | Mean of completed resolution intervals across a defined period and population | End of the reporting period | Average restoration duration | Tail performance or unreported faults |
| Availability | Operable time divided by scheduled time | End of the reporting period | Serviceable time over a stated denominator | Whether the selected population reflects business need |
| First-time fix | A completed visit | No repeat visit for the same symptom within the stated rule | Whether a visit ended without an immediate repeat | Whether the root cause was understood |
The most important split is between operational restoration and administrative closure. A door can be operable while a ticket remains open for notes, evidence, or a replacement process. Resolution therefore stops at verified operability; ticket closure remains a separate service-process timestamp.
The Uptime and MTTR KPIs Worksheet: What to Write Down and Why turns these boundaries into record fields. Use it to test whether two teams would calculate the same interval from the same evidence.
Set the population, schedule, and clock
The population must be explicit. It may be all installed doors, doors at one site, or a defined operational subset. The unit identifier should come from the object schedule rather than an informal room label, so the same door remains traceable when rooms are renamed or occupancy changes. The Office Smart Lock Site Survey and Smart Lock Retrofit Site Survey provide the type of door-by-door context needed to keep that schedule controlled.
Scheduled time also needs a policy. State which periods the door was expected to be available, how planned maintenance is treated, and what happens when a door leaves service. Preserve the original scheduled value and record any approved adjustment separately. Do not silently remove difficult periods from the denominator.
Clock ownership must be equally explicit:
- Use one reporting time zone or show converted timestamps beside local entries.
- Synchronize operational systems and record the source of each critical timestamp.
- Define which channel starts the fault clock when a report arrives by more than one route.
- Keep raw timestamps intact and apply display rounding only when producing the report.
- Name the dataset, such as
uptime-and-mttr-kpis, so filters and exports cannot be mixed accidentally.
Assign an authoritative source to every event
A service-ticket workflow, device feed, and technician note can all contribute, but they should not compete for the same field. The MTTR measurement procedure begins with an event ownership map that identifies the source of truth and the evidence used to challenge it.
| Event | Authoritative source | Supporting evidence | Evidence that should not stand alone |
|---|---|---|---|
| Fault reported | Agreed intake channel and its timestamp | Caller log, email, scanner, or portal entry | A later ticket-creation time that hides the earlier report |
| Acknowledged | Service-desk or operator log | Agent note and status history | A generic status change without an actor |
| Attendance | Technician attendance entry | Work record or door-visit note | A remote app action that may not represent attendance |
| Door operable again | Door-level verification and test result | Photo, reading, or supporting observation | Ticket closure alone |
| Ticket closed | Workflow status change | Resolution note and linked evidence | Operability verification |
| Repeat event | Link to the prior episode | New symptom and inspection record | A new ticket with no relationship to the earlier fault |
Service ticket timestamps remain useful because they create a searchable operational history, but they need a recorded reconciliation rule. Telemetry is useful when the deployed system exposes a governed event feed, yet an inferred state change should not overwrite a direct door test. Manual-only recording can work, but it requires disciplined completion and later audit. The best architecture is the one that meets the evidence requirement without creating records the team will not maintain.
Trace the event chain from report to restoration
A readable timeline prevents the common error of treating different service events as interchangeable. The chain below shows the order used for the KPI record.
flowchart TD
A[Fault reported] --> B[Report time recorded]
B --> C[Acknowledgement recorded]
C --> D[Attendance recorded]
D --> E[Door operability verified]
E --> F[Resolution time closed]
F --> G[Evidence linked]
G --> H[Period KPI calculated]
The downtime interval record begins with the report time and ends with verified operability. Acknowledgement, attendance, parts dispatch, and administrative closure remain visible intermediate fields. They explain the restoration path without changing the resolution boundary.
Record restoration when the door satisfies the operating requirement defined for that site. That requirement may include a prescribed access method when normal occupants depend on electronic credentials. If the agreed fallback route is sufficient for that door, the facility manager should say so in advance rather than changing the interpretation after an incident.
A repeat symptom can start a new resolution interval, but the record should preserve the link to the earlier episode. The reporting rule must distinguish a genuinely new failure from the continuation of an unresolved incident. Never close one interval and conceal the next behind the same ticket number.
Capture the minimum evidence at the door
The field service KPI log must contain enough information to reconstruct the calculation and challenge the conclusion. A list of actions without the reason for each action is not evidence.
| Record group | Required fields | Audit question |
|---|---|---|
| Object identity | Door or unit identifier, model, and variant | Can the asset be located without relying on a room nickname? |
| Event times | Reported, acknowledged, attended, operable again, and administratively closed | Does every interval use the agreed boundary? |
| Symptom | Technician’s description and measured observation | What was seen rather than inferred? |
| Checks | Checks performed, order, and what each check ruled out | Does the record show why the technician moved to the next step? |
| Action | Parts fitted, unit swapped, and disposition of the removed unit | What changed at the door? |
| Credential state | State before and after work where credentials were involved | Was the access layer checked and restored? |
| Evidence | Photo, reading, test result, or other supporting record | Can a reviewer inspect the conclusion? |
| Classification | Severity, suspected cause, incident type, and repeat link | Has an event been reclassified without an audit trail? |
| Data provenance | Reporter, recorder, source system, and later adjustments | Who changed the record, when, and why? |
The model and variant should be transcribed from the asset label and controlled catalogue. When a model-specific fact affects a support or procurement decision, confirm on the dated product page for the specific SKU. Do not infer a model capability from a similar unit.
A field record without the “what each check ruled out” field remains an activity list. That column connects the technician’s observation to the action taken and makes later root-cause work possible.
Calculate availability from operable and scheduled time
The smart-lock uptime calculation should use a simple denominator:
Portfolio availability = total operable door-time / total scheduled door-time
A door contributes operable time only while it satisfies its defined operating requirement. Scheduled time comes from the approved service calendar, not from a ticket system’s activity. A door that receives no ticket may still have been unavailable, so the absence of a ticket must never be treated automatically as evidence of availability.
Aggregate at the level the decision requires. Sum operable and scheduled time across the selected doors when calculating a portfolio measure. An unweighted average of door percentages gives a small door the same influence as a high-use door, so it should be labelled separately and not mixed with the aggregate result.
Use these publication controls:
- State the population, period, time zone, and treatment of planned maintenance.
- Report unresolved state as unknown rather than assigning an undocumented value.
- Keep data completeness visible beside the operational measure.
- Stop the downtime period at verified operability, not at ticket closure.
- Preserve every denominator adjustment with its reason and approval.
This process produces a result that can be explained, even when an event was discovered late or a record had to be reconciled.
Calculate MTTR from completed restoration intervals
MTTR is the mean of resolution intervals across a defined period and a defined population:
MTTR = sum of completed resolution intervals / count of completed fault episodes
Each episode must have a report time and a verified operability time. Open incidents at the end of the period do not acquire an invented resolution time merely because the reporting period closed. Report their count and aging separately; include them in MTTR only after restoration has occurred and the episode enters the defined calculation cohort.
The mean can conceal a long tail. Retain the mean as the stated KPI, but accompany it with measures such as the distribution, an upper-tail measure, and the longest open or completed interval as the buyer’s policy requires. These are supporting views, not replacements for MTTR.
Keep response and attendance intervals beside the result so reviewers can see where time accumulated. First-time fix and repeat rate also require their own stated windows. Do not combine them into MTTR unless the metric definition explicitly says to do so. A low mean can coexist with repeated failures, while a single prolonged restoration can materially affect the mean; the KPI pack should expose both conditions.
Reconcile conflicting records
Reconciliation should preserve the original event and the reason for any adjustment. It should not create a second hidden dataset.
| Conflict | Reconciliation test | Required record |
|---|---|---|
| Ticket says closed, but the door was not operable | Review the verification evidence and reopen or adjust the event | Closure status and operability kept separate |
| Device state differs from the door test | Establish which service requirement defines operability | Direct observation and inferred state both retained |
| The same fault appears on several tickets | Link the episodes and select the authoritative report event | Episode and root-cause relationships |
| Ticket was created after the fault was reported | Compare the agreed intake channel with creation time | Earlier report time plus late-entry flag |
| A fault is found during inspection without a prior report | Record discovery and the approved availability treatment | Discovered time, reported time, and policy reference |
| An outage is still open at period close | Exclude it from completed MTTR and report it as open | Aging and next review action |
The facility manager should approve any rule that backdates an event or changes a denominator. The record must show the value before the change, the value after it, the person responsible, and the evidence used. A blank state, “unknown,” and an operational zero are three different entries and must never share a field.
Choose the reporting view and treat edge cases consistently
A lock availability record can support several views, but each view needs its own label and denominator. Do not place site averages, portfolio aggregates, and single-door results under one unlabeled KPI.
| Reporting view | Suitable question | Required rule | Main caveat |
|---|---|---|---|
| Portfolio availability | How much of the selected estate was operable? | Aggregate operable and scheduled door-time | Population and exclusions must be visible |
| Site availability | Which site needs operational attention? | Aggregate doors within that site | Site operating schedules may differ |
| Critical-door view | Are designated doors meeting the stated service requirement? | Use the approved subset and its schedule | It is not a portfolio replacement |
| Door record | What happened at this asset? | Preserve every event and evidence link | A single door can distort an average |
| MTTR distribution | Where did restoration time accumulate? | Report completed episodes and open incidents separately | The mean alone does not show the tail |
| Data-quality view | Can the result be reproduced? | Show missing, late, adjusted, and unknown records | Completeness is not service performance |
Use the Running Uptime and MTTR KPIs Across Several Sites guidance when site schedules, populations, or data ownership differ. Edge cases should follow an approved decision tree rather than become informal exceptions. The Uptime and MTTR KPIs: Which Branch Do You Take can route late reports, duplicate tickets, repeat symptoms, and end-of-period open incidents into consistent branches.
Audit the KPI record before publication
At each reporting cycle, trace a representative selection back to the underlying events. Include doors with unusual restoration intervals, records adjusted after closure, incidents discovered late, and cases where several methods disagree. The sample should challenge the result rather than merely repeat it.
The audit should ask:
- Does the report time match the agreed intake channel?
- Was operability verified at the door or only inferred from a software state?
- Does each check state what it ruled out?
- Are response, attendance, resolution, and closure stored as separate fields?
- Is an open incident excluded from completed MTTR but visible in the open-incident report?
- Has any denominator or severity rule changed since the prior period?
- Can a reviewer reproduce the published value from the retained record?
This is where uptime and mttr kpis measurement becomes an operating routine rather than a reporting task. Every published value must be traceable to an event record, and every correction must remain visible. Record data-quality exceptions beside the KPI rather than burying them in an explanatory note that is easy to miss.
The practical counterpart to this point is smart lock retrofit for apartment buildings. That is the same question uptime and mttr kpis measurement has to answer here.
Select the operating model and implement it in sequence
The final choice depends on the evidence the buyer can maintain, not on the sophistication of the dashboard. Use the decision table below as the starting point, then adjust for the operating controls already in place.
| Operating condition | Preferred approach | Reason | Condition to revisit |
|---|---|---|---|
| Small controlled estate with a consistent field routine | Manual field service KPI log plus ticket summary | Direct verification with limited systems overhead | Manual completion becomes unreliable |
| Estate already managed through a service desk | Ticket spine with door-verification fields | Preserves workflow accountability and the restoration event | Late or incomplete ticket updates persist |
| Deployed devices expose a governed event feed | Telemetry-supported hybrid | Adds observations while retaining direct verification | Synchronization gaps undermine event fidelity |
| Multi-site portfolio with mixed contractors | Reconciled hybrid | Makes the event source and population explicit | Sites cannot maintain common definitions |
Implement the method in a deliberate sequence:
- Approve the population, service requirement, event boundaries, and exclusions.
- Configure the event dictionary, source ownership, and audit trail.
- Test records through a representative pilot without publishing the result as a performance claim.
- Reconcile open incidents and historical records according to the approved policy.
- Run the chosen method in parallel with any existing report and explain differences.
- Publish the result with population, denominator, data-quality, and open-incident notes.
- Review the definitions and source map whenever the estate, service workflow, or ownership changes.
The Uptime and MTTR KPIs for Installed Smart Locks: A Buyer and Operator Guide supports the briefing of facility and service teams around these rules. Before the next service review, assign a KPI owner, approve the event dictionary, and require one reproducible path from every reported fault to the record that proves the door is operable again.
Put the answer on paper and request a quote against that document.

