Skip to content

property management operations research

Turnover Days Are Not One Number: A Vacancy Reporting Data Quality Review

A data-quality review of turnover and vacancy reporting for rental portfolios, with event definitions, denominators, reconciliation checks, and honest limits.

| 6 sources

Turnover Days Are Not One Number: A Vacancy Reporting Data Quality Review
Table of contents

Research question and scope

This brief asks how a property management team can report turnover and vacancy without collapsing several different intervals into one misleading number. It does not ask whether a particular vacancy is acceptable. Acceptability depends on the owner's budget, the local market, the condition of the unit, and the lease terms. Those are management decisions, not measurement definitions.

The unit of analysis is one vacant unit episode. An episode begins when a unit becomes vacant under the team's written definition and ends when it is occupied again or removed from the rent-ready pool. A single episode may include a move-out inspection, cleaning, repairs, listing, showings, an application, a lease signature, and a move-in. Each of those is a milestone, not a separate episode.

The method is descriptive. It can show where dates conflict, which interval is being reported, and whether the same episode is counted twice. It cannot prove that a delay was avoidable, that a vendor was slow, or that a marketing decision caused a longer vacancy. Those claims need a different design and better evidence.

Why a single turnover number fails

"Average turnover days" sounds like one fact. In practice it is at least five different intervals. The first is the time from move-out to the unit being physically ready. The second is the time from ready to listed. The third is the time from listed to an approved application. The fourth is the time from approval to lease start. The fifth is the time from lease start to actual occupancy, which may differ if the resident moves in late.

A team can improve one interval and worsen another. Faster cleaning may not help if the unit sits unlisted for a week. A shorter listing period may not help if the applicant cannot sign for two weeks. Reporting only the total hides which stage deserves attention.

The total is also sensitive to the start and end definitions. If the episode starts at lease end rather than key return, a resident who stays three extra days adds three days to turnover. If the episode ends at lease signature rather than move-in, a unit can look occupied while it is still empty. Neither choice is wrong, but the report must state which one it uses.

Finally, averages hide the shape of the data. A portfolio with many quick turns and a few long repairs can have a reasonable average and a terrible tail. The median and the distribution are usually more honest than the mean.

Operational definitions

A reporting dictionary should define each event before anyone calculates a rate. The following definitions are a working example, not an industry standard.

Vacancy start. The date the unit becomes available for turnover under the team's rule. Common choices are lease end, key return, or the date the resident's possession ends. Pick one and use it for every property in the report.

Possession end. The date the resident no longer has the right to occupy the unit. This may differ from key return. Record both when they differ.

Unit ready. The date the unit passes the team's make-ready checklist. The checklist should name the required inspections, cleaning, repairs, and safety checks. A unit is not ready because a vendor says so; it is ready when the checklist is complete.

Listed. The date the unit first appears as available to prospective residents. If the listing is paused and restarted, record the first live date and the restart date separately.

Application approved. The date the team approves an applicant under its written criteria. Approval is not a lease.

Lease signed. The date the last required signature is received. A verbal commitment is not a signature.

Occupied. The date the resident takes possession under the lease. This may be the lease start date or a later agreed date.

Episode close. The date the unit is occupied, removed from the rental pool, or converted to another use. Record the reason. A unit taken off the market is not the same as a unit rented.

The dictionary should also define exclusions. A unit held for an owner's renovation, a unit damaged by a fire, and a unit in a legal hold may need to be reported separately rather than mixed into routine turnover.

An event model for a vacant unit

A reviewer should be able to reconstruct an episode as a sequence. The table is a design aid, not a compliance checklist.

Stage Question Evidence
Vacancy start When did the unit become available? Lease end, notice, key return, possession record
Condition What condition was the unit in? Move-out inspection, photos, work orders
Make-ready What work was required and when was it finished? Checklist, vendor invoices, completion photos
Listing When did the unit go live and at what terms? Listing record, portal export, photo set
Demand How many inquiries and showings occurred? CRM log, showing notes, application count
Screening When was an application received and decided? Application record, approval note
Lease When was the lease signed and when does it start? Signed lease, signature timestamps
Occupancy When did the resident take possession? Move-in record, key log, utility transfer

The sequence exposes gaps that a single number cannot. If the unit was ready on the fifth but not listed until the twentieth, the make-ready interval looks fine and the listing interval looks bad. If the lease was signed on the first but occupancy began on the fifteenth, the unit was economically committed but physically empty. Both facts matter to different readers.

The model also forces a decision about overlapping stages. Cleaning and repairs may overlap. Listing may begin before the unit is fully ready, with a note that showings start later. The report should state whether overlapping stages are allowed and how they are counted.

Measures and denominators

A small set of measures is more useful than a composite score.

Total vacancy days. The number of days from vacancy start to occupancy, for each closed episode. Report the median and interquartile range, not only the mean. Also report the count of episodes and the number excluded.

Make-ready days. Days from vacancy start to unit ready. This interval is closest to the work the maintenance and turnover team controls.

Listing lag. Days from unit ready to first live listing. A positive lag means the unit was ready but not marketed.

Days on market. Days from first live listing to approved application. State whether paused days are included.

Approval-to-signature days. Days from application approval to the last required signature. This interval is sensitive to applicant responsiveness and screening complexity.

Signature-to-occupancy days. Days from lease signature to actual possession. A long interval may be a planned future start date rather than a delay.

Rent-ready rate. Closed episodes that ended in occupancy, divided by all closed episodes in the period. Report cancellations and owner holds separately. A high rent-ready rate can coexist with long vacancy if the units that do rent take a long time.

Each measure describes a record property. None of them proves that a delay was avoidable. A unit can be ready quickly and still sit vacant because the market is soft or the asking rent is above what applicants will pay. The measures show where to look, not what to conclude.

Data quality checks, limitations, and reporting protocol

Before calculating anything, test the episode records for impossible or ambiguous states. Look for a unit ready date before the vacancy start, a lease signed before the application was approved, an occupancy date before the lease start without an explanation, and a listing date after the first showing. Check whether a blank date means "not applicable" or "not recorded." Those are different facts.

Check for duplicate episodes. A unit may be entered as vacant when the resident gives notice, then entered again when keys are returned. If both records are counted, the portfolio appears to have more vacancies and a shorter average. A merge rule should be written before the review.

Check whether the same unit appears in two property records. This is common after a portfolio transfer or a unit renumbering. Compare the address, unit number, and lease history rather than relying on the property name alone.

Check the listing export against the property system. A listing that was removed and relisted may appear as two listings with different start dates. The report should state which one is used.

The method has limits. It relies on records the team chose to create. It cannot observe showings that were never logged, applicant conversations that were never recorded, or owner decisions that were made by phone. A short vacancy may reflect a strong market rather than an efficient process.

Local definitions of vacancy, notice, and possession vary. The sources cited below provide context for housing markets and records practice, not a jurisdiction-specific rule. The manager should confirm applicable lease terms and local requirements with counsel.

Finally, the review cannot establish causation. A portfolio with long turnover may have older buildings, more complex repairs, or a seasonal leasing pattern. Those differences should be described, not treated as performance effects.

A practical protocol has four steps. First, write the event dictionary and keep it stable for at least a quarter. Changing definitions mid-period makes comparison impossible. Second, assign one owner for the episode record. The coordinator can maintain dates and evidence; the manager approves the ready status and any exception. Third, reconcile the episode list against the rent roll and the listing export before publishing. Fourth, report the median, the interquartile range, and the count of excluded episodes alongside the total.

The summary should lead with scope: how many episodes, which period, which properties, and which exclusions. Present the interval measures with numerators and denominators. Add a table of data defects, such as missing ready dates or duplicate episodes. Close with the one or two record changes the team will make and the date of the next review.

Turnover is not one number. It is a chain of events, each with its own owner and evidence. A team that reports the chain honestly can see where a unit actually waited, and can improve the stage that deserves attention instead of arguing about an average that hides the answer.

Sources