How to Calculate Void Periods in Coliving (And What a Good Rate Looks Like)

Void is the number every investor asks about and the number most shared-living operators calculate slightly wrong. Not because the arithmetic is hard, but because the unit of measurement is wrong. In a 6-bed HMO, a per-unit void rate can read 0% while a third of your revenue is missing.

This post covers the formulas, the definitions that need pinning down before the formulas mean anything, worked examples you can copy into a spreadsheet, and a method for setting a void target you can actually defend in a board meeting.

Void nights, not void days

Start with the atomic unit. Rent accrues nightly, so void should be counted in nights.

Void nights for a single letting gap:

Void nights = (start date of next agreement) - (end date of previous agreement)

Where "end date of previous agreement" is the last night paid for, not the day the resident physically moved out.

That distinction matters more than it sounds. If a resident's agreement runs to 31 March but they hand back keys on the 20th, and the next resident starts on 25 March, you do not have void. You have five nights of overlap, which is five nights of double-paid revenue. Book it as such. Operators who measure void from the physical move-out date systematically overstate void and understate the value of their notice terms.

Bed-night occupancy: the base formula

Everything else derives from this.

Available bed-nights = beds in service x nights in period
Occupied bed-nights  = sum of nights each bed was under a paid agreement
Occupancy            = Occupied bed-nights / Available bed-nights
Void rate            = 1 - Occupancy

Note "beds in service". A bed sealed off for a six-month refurbishment is not a letting failure, and burying it in your void rate destroys the signal you need from the number. More on that below.

The two-lever decomposition

The single most useful thing you can do with void is stop treating it as one number. Void rate is the product of how often beds turn over and how long each gap lasts:

Annual void nights per bed = (turns per bed per year) x (average void nights per turn)
Annual void rate           = Annual void nights per bed / 365

And turns per bed per year is just:

Turns per bed per year = 12 / (average tenancy length in months)

Two levers, two different teams. Turnover frequency is owned by community, pricing and renewals. Void nights per turn is owned by lettings and maintenance. If void is rising and you only report the headline, you have no idea which team to talk to.

Worked example

Suppose a 120-bed portfolio, all beds in service, with these illustrative inputs:

• Average tenancy length: 9 months
• Average void nights per turn: 18

Turns per bed per year     = 12 / 9 = 1.33
Annual void nights per bed = 1.33 x 18 = 24 nights
Annual void rate           = 24 / 365 = 6.6%
Occupancy                  = 93.4%
Void bed-nights (portfolio)= 120 x 24 = 2,880

At an illustrative achieved rate of £30 per bed-night, that is £86,400 of lost revenue a year.

Now change one lever at a time. Cutting void nights per turn from 18 to 10 takes annual void nights per bed to 13.3 and the void rate to 3.6%. Extending average tenancy from 9 to 12 months, leaving void nights per turn at 18, takes it to 18 nights and 4.9%. Same headline improvement, completely different projects.

Sensitivity table

Annual void rate at combinations of average tenancy length and void nights per turn:

Read across a row and you can see why short-stay shared living is so unforgiving. At a 6-month average stay, every extra week of turnaround costs you nearly four points of occupancy.

Per-bed void is not per-unit void

This is the part that catches out anyone who has come from a single-family or BTR background.

Per-unit void asks: is this property let? Per-bed void asks: how much of this property is earning? In shared living, the first question is close to useless.

Take three illustrative properties over a 30-day month:

Per-unit void says one property in three is empty. Per-bed void says well over half your inventory is not earning. Only one of those numbers connects to your bank balance.

There is a second, subtler trap in that table. The portfolio per-bed void of 58.3% is calculated from the bed-night totals (315 / 540), not by averaging the three per-bed percentages. Averaging the percentages gives 48.6%, which is wrong, because it weights a 4-bed house the same as an 8-bed house. Always aggregate the numerator and denominator, then divide. Never average rates.

The one place per-unit void still earns its keep is whole-building risk: a fully empty building carries standing costs, insurance conditions and council tax exposure that a scattering of empty beds does not. Report it as a separate risk metric, not as your occupancy metric.

Five ways operators accidentally hide void

1. Snapshot occupancy instead of bed-night occupancy. If you measure occupancy on the first of each month, a bed that is empty from the 8th to the 26th shows as fully occupied in both snapshots. Every mid-month gap disappears. Snapshot occupancy is fine for a board slide; it must never be your operating metric.

2. Averaging percentages across properties. Covered above, and it is astonishingly common in spreadsheet reporting because the AVERAGE function is right there next to the numbers.

3. Treating out-of-service beds as void, or ignoring them entirely. Both are wrong. A bed out for refurbishment is a capital decision, not a lettings failure, so leaving it in your void rate makes your lettings team look bad and hides real problems. But removing it entirely hides genuine revenue loss and lets a "temporary" closure run for eight months unnoticed. Report both:

• Lettable void rate = void bed-nights / bed-nights in service. This is your lettings performance number.
• Total revenue-loss rate = (void + out-of-service bed-nights) / total physical bed-nights. This is the number the P&L feels.

Track the gap between them. If it widens, you have beds quietly out of service that nobody is chasing.

4. Excluding reserved and pre-let beds from the denominator. A bed that is signed for a move-in three weeks out is still not earning tonight. It belongs in available bed-nights and in void bed-nights until the paid term starts. Pulling it out of the denominator is how a pipeline gets counted as revenue.

5. Mixing structural and frictional void. A bed that has sat empty for 200 days is not the same problem as twenty beds with a 10-day turnaround. Split them, because the fixes are different.

Structural vs frictional void

Frictional void is the unavoidable gap created by turnover. Structural void is inventory that repeatedly fails to let: the box room, the ground-floor front bedroom next to the bins, the bed in a house with a housemate dynamic that pushes people out.

A practical split: any turn where void nights exceed roughly three times your median is a candidate for structural review, not a lettings statistic. Pull those beds out, list them individually, and put a named owner and a decision against each: reprice, refurbish, reconfigure, or accept and plan around it.

Once you separate them, the reporting gets honest. Frictional void responds to process. Structural void responds only to price or capital.

Decomposing void nights per turn

To manage frictional void you need to break the gap into its parts. Every void gap is some combination of:

• Turnaround nights: last paid night to bed ready to occupy. Clean, repair, inspect, replace linen, re-key.
• Marketing nights: bed ready to occupy, no signed agreement yet.
• Lead-in nights: agreement signed, resident's start date is still in the future.

These overlap in practice, and that overlap is the whole game. If you receive a month's notice, marketing can start the day notice lands, and turnaround can happen in the last few days of the paid term. Under those conditions your theoretical floor for void nights per turn is close to your turnaround SLA alone.

So the honest test is not "how does my void compare to the market". It is:

Excess void nights per turn = actual void nights per turn - turnaround SLA

If your turnaround SLA is 3 nights and you are running 18 void nights per turn, 15 of those nights are a lettings and notice-management problem, not a cleaning problem. That is a number a team can act on this quarter.

So what is a good void rate?

Nobody should give you a single number, and you should distrust anyone who does. Void rates are not comparable across operators without normalising for at least these five things:

1. Stay length. A 3-month-average flex product and an 18-month HMO have structurally different void floors.
2. Seasonality and cycle. PBSA and most student housing run on a September cycle, so annual void rates are dominated by summer. Measure those against the academic year, or against the same weeks last year, never against a rolling 12-month portfolio average.
3. Notice terms. One month's notice versus two weeks changes your achievable floor before anyone does any work.
4. In-service definition. Operators who quietly park difficult beds as "out of service" post better void rates for free.
5. Whether reserved beds sit in the denominator.

The useful benchmark is your own distribution. Take the last 12 months of turns, and report:

• Median void nights per turn (your typical performance)
• 90th percentile void nights per turn (your tail, where the money leaks)
• Turnaround SLA (your theoretical floor)
• The gap between median and floor (your addressable opportunity)

Then set the target as a movement: cut the 90th percentile by a third, or close half the gap between median and floor. That is defensible, it is attributable to a team, and unlike a borrowed industry figure it is actually true of your portfolio.

Where this connects

Void is one input into revenue per available bed, which is the metric that ties occupancy and pricing together. See the companion post on calculating and benchmarking per-bed yield, the wider set of per-bed reporting KPIs, and why per-bed inventory beats per-room inventory for the data model that makes any of this calculable. Definitions for void period, RevPAB and per-bed inventory are in the shared living glossary.

Calculating void this way by hand is possible but tedious, which is why most operators only do it once a quarter. Bed-night level occupancy, turn-by-turn void decomposition and the in-service split are all standard views in JumboTiger's inventory and occupancy module.