Maintenance Response Time Measurement: A Design That Survives Scrutiny

Response time is one of the most quoted and least defined measures in rental operations. A team reports a four-hour average and a resident describes three days of silence. Both can be true because they are measuring different events. Before a portfolio uses response time as an operating signal, it needs to specify which clock it is reading and what stops that clock.

This research note proposes a four-clock model, sets out the data each clock requires, and explains where the measure can mislead. It is a measurement design, not a benchmark. No universal target is proposed because consequence, season, jurisdiction, and property type change what a reasonable interval is.

Why a Single Average Fails

A single average blends at least four different intervals:

• From the resident's message to the operation's receipt.
• From receipt to a human triage decision.
• From the decision to vendor dispatch or scheduling.
• From dispatch to verified completion.

A fast acknowledgement and a slow repair can produce the same headline number as a slow acknowledgement and a fast repair. The first looks responsive and feels negligent. The second looks unresponsive and may be perfectly managed. Averaging them together produces a figure that supports no decision.

Averages also hide the tail. Maintenance queues are usually skewed: most items resolve within the service standard and a small number take far longer because of parts, access, approvals, or missed follow-up. The mean moves slowly while the tail determines resident experience and risk.

Define the Four Clocks

Clock 1: Receipt. Starts when the resident sends the message through an approved channel. Stops when the request enters the operating record. This clock measures channel hygiene: shared inboxes, after-hours coverage, and message routing.

Clock 2: Triage. Starts at receipt. Stops when a person records a consequence category, an owner, and a next action. This clock measures decision speed, not repair speed.

Clock 3: Dispatch or scheduling. Starts at triage. Stops when a vendor or internal technician has a confirmed window, or when the item is placed in a scheduled routine batch with a date. This clock measures coordination.

Clock 4: Resolution. Starts at dispatch. Stops when the repair is verified, the resident is informed, and the record is closed. This clock measures delivery and evidence.

Each clock has a different owner and a different failure mode. Reporting them separately makes the queue legible without pretending that one number explains the whole service.

Instrument the Transitions, Not the Statuses

Most systems record a status such as "in progress" but not the moment of transition. A timestamp at each transition is what makes the clocks computable. At minimum, record:

• Received at
• Triaged at
• Scheduled at, with the confirmed window
• Completed at, with verification
• Closed at

Where a request moves backwards, record the rework reason. A reopened item is not a data error. It is the most useful signal in the dataset.

Handle After-Hours and Channel Gaps

A resident who calls at 22:00 and reaches voicemail has not had their request received. If the operation counts receipt at the next morning's inbox check, Clock 1 is honest but the resident experience is not. Report both the time of the resident's attempt and the time of receipt so the gap is visible.

The same applies to messages sent to an individual rather than the approved channel. These are common, and excluding them makes the operation look better than it is. Classify them separately rather than dropping them.

Separate Consequence From Speed

A sewage backup and a dripping tap should not be judged against the same interval. Attach the triage category to every record and report clocks by category. A portfolio may accept a longer interval for cosmetic work and a very short interval for safety items. The categories should be defined in advance and applied consistently.

Where a legal or contractual standard applies, record the standard and its source. Do not invent a universal deadline from a local rule.

Control for Access and Resident-Caused Delay

Many long intervals are not the vendor's fault. The resident may be unavailable, the unit may be inaccessible, or a required approval may sit with an owner. These are legitimate stops, but they need to be recorded as pauses with a reason rather than absorbed silently into the average.

A practical approach is to report both elapsed time and active time. Elapsed time is what the resident experiences. Active time is what the operation controls. Reporting both prevents two errors: blaming the team for resident delay and hiding operational delay behind it.

Sample and Audit

Automated timestamps are only as good as the behavior behind them. Once a month, sample ten closed items and compare the record to the message thread, the vendor confirmation, and the completion evidence. Look for:

• Requests marked triaged with no recorded decision
• Scheduled items with no confirmed window
• Completions without verification
• Closures that took place before the resident was informed

This audit is short and catches the gap between a well-designed field and a well-used one.

Publish the Definition With the Number

Any response-time report should state the clocks, the start and stop events, the treatment of pauses, the categories, and the sample period. A number without a definition invites comparison with other portfolios that measure something else.

Used this way, response time becomes a diagnostic rather than a scoreboard. It shows where work waits: at receipt, at decision, at scheduling, or at verification. That is the information a portfolio can act on.

Sources and Method Notes

This design draws on general quality-management practice and published housing operations guidance. For program context, see the U.S. Department of Housing and Urban Development and HUD User research. For household and housing condition data that can inform baselines, see the American Housing Survey. For consumer-facing rental guidance, see the Consumer Financial Protection Bureau. The four-clock model and the audit checklist are the author's proposed method, not a published standard.