Per-Bed Yield: The Metric Shared-Living Operators Should Report On

Ask five shared-living operators for their per-bed yield and you will get five different numbers, computed five different ways, none of which are comparable. That is not sloppiness. "Yield" genuinely refers to at least three distinct metrics, aimed at three different audiences, and most reporting packs mix them.

This post separates them, gives the arithmetic for each, and then deals with the part that actually causes bad decisions: the definitional choices in the numerator and the denominator that can reorder your entire portfolio ranking without changing a single letting.

Three metrics, three audiences

They stack. RevPAB feeds NOI per bed, which feeds net yield per bed. If the first is wrong the other two are wrong, which is why operators should own RevPAB with some care rather than treating it as a finance metric.

RevPAB, and why it is two metrics in disguise

RevPAB  = Rental revenue / Available bed-nights
RevPOB  = Rental revenue / Occupied bed-nights     (your achieved rate)
RevPAB  = RevPOB x Occupancy

That last identity is the useful one. RevPAB moving is never self-explanatory. It moved because your achieved rate moved, or because occupancy moved, and those are different problems with different owners. Report all three or the number tells you nothing actionable. Occupancy in this identity is bed-night occupancy, which is covered in detail in our post on calculating void periods.

Use bed-nights, not bed-months. If you compute monthly RevPAB from a calendar month, February will always look weak and March strong. At a constant £25 per bed-night, February produces £700 per bed and March produces £775, an apparent 10.7% improvement caused entirely by the calendar. Bed-nights remove that noise. Convert to a monthly figure at the end if your board prefers it, using a consistent 30-night month.

Why per-unit metrics mislead in shared living

Take three illustrative properties, one letting unit each, over a 30-night month.

Per unit, the ranking is Ashby, Nelson, Church. Per bed, it is exactly inverted: Church, Nelson, Ashby. Ashby Road generates the most cash and the least value per unit of inventory. Any capital allocation decision made on the per-unit column is being made backwards.

This is the mix effect, and it is not a rounding issue. Per-unit metrics are only comparable when units are comparable, and in shared living they are structurally not. A portfolio with 4-bed, 6-bed and 10-bed configurations has no meaningful "average rent per property".

There is a second failure. Per-unit occupancy cannot detect partial vacancy. A 6-bed house with two empty rooms is 100% occupied at unit level and 67% at bed level. In a portfolio measured per unit, a third of your revenue can vanish while occupancy reads perfect.

Now add occupancy

Same three properties, decomposed:

Nelson House has the highest achieved rate in the portfolio and still ranks below Church Street on RevPAB, because it is not filling. Ashby Road is full and still ranks last, because it is underpriced or over-bedded. Three properties, three completely different interventions, all invisible in a per-unit report.

The numerator problem: what counts as revenue

This is where cross-portfolio and cross-operator comparisons quietly break.

Continue the example. Suppose Ashby Road and Church Street are bills-excluded, while Nelson House is bills-inclusive and carries an illustrative £110 per bed per month of utility cost inside its rent.

The ranking changes again. Nelson House drops from second to third purely because of a billing convention. Across a real portfolio with a mix of inclusive and exclusive products, an unnormalised RevPAB comparison is close to meaningless.

A workable rule: define net rental RevPAB as rent only, with pass-through utilities, service charges and ancillaries reported as separate lines. Then:

• Include in net rental RevPAB: contractual rent, rent-equivalent room upgrade premiums, and any recurring charge the resident cannot decline.
• Report separately: utilities and bills-inclusive components, parking, storage, laundry, cleaning add-ons, pet fees, and anything optional.
• Exclude entirely: deposits, refundable holding fees, and any amount you are holding rather than earning.
• Decide once and document it: late fees, dilapidation recharges and early-termination fees. There is no universally right answer. There is only a consistent one.

Whatever you choose, write it down in a one-page metric definition and put it in front of anyone who receives the report. The single biggest cause of "the numbers changed" arguments is an undocumented numerator.

The denominator problem: what counts as available

Available bed-nights sounds objective. It is not.

• Physical beds vs lettable beds. A room licensed for single occupancy but containing a double bed is one bed, not two. Count contractually lettable beds.
• Beds out of service. A bed under refurbishment is not available. Excluding it raises your RevPAB, which is legitimate for measuring lettings performance and misleading for measuring asset productivity. Report both: RevPAB on in-service beds, and RevPAB on total physical beds. The gap is your capital drag.
• Reserved and pre-let beds. These are available. A signed agreement starting in three weeks earns nothing tonight.
• Beds added or removed mid-period. Prorate them into available bed-nights from the date they entered or left service. Counting a bed that existed for nine days of a month as a full bed understates RevPAB.

Operators who quietly park difficult inventory as "out of service" post better RevPAB for free. If you are benchmarking against anyone else's published figure, this is the first thing to ask about, and you will usually not get a straight answer.

From RevPAB to NOI per bed

Once RevPAB is defined properly, the cost side is comparatively simple, provided you allocate costs to beds rather than to properties.

Opex per available bed = Total operating cost / Available beds
NOI per bed            = Annualised RevPAB - Opex per available bed

Costs that should be allocated per bed rather than per property include cleaning and turnaround, resident-facing staff time, utilities on inclusive products, wear-and-tear R&M, and letting and marketing costs. Costs that are genuinely property-level, insurance, licensing, building compliance, structural R&M, should be allocated across the property's beds so that the per-bed figure is complete, but tracked separately so you can see which is which.

An illustrative annualisation for Church Street:

Annualised RevPAB   = £800 x 12          = £9,600 per available bed
Opex per bed        =                      £3,100   (illustrative)
NOI per bed         = 9,600 - 3,100      = £6,500

Net yield per bed, for the investor conversation

Gross yield per bed = Annual gross rent per bed / Capital value per bed
Net yield per bed   = NOI per bed / Capital value per bed

Continuing the illustration at a capital value of £95,000 per bed:

Gross yield = 9,600 / 95,000 = 10.1%
Net yield   = 6,500 / 95,000 =  6.8%

The gap between those two numbers is the operating business. In single-family lettings that gap is narrow. In shared living it is wide, because shared living is an operating business wearing a real estate costume. Reporting gross yield per bed to an investor without the net figure alongside it overstates the asset and understates your own value.

How to benchmark without fooling yourself

External benchmarks in shared living are mostly not comparable, for all the reasons above plus stay length, market, product tier and vintage. Two practical approaches that do work:

1. Benchmark internally, in cohorts. Group beds that are genuinely alike, then compare within the group. Useful cohort dimensions:

• Market or micro-location
• Product tier (ensuite, standard, studio)
• Property vintage or refurbishment year
• Configuration size (4 to 6 beds, 7 to 10 beds, 10+)
• Stay-length profile

Within a cohort, RevPAB dispersion is a real signal. A bed running 20% below its cohort median is either mispriced, in poor condition, or in a house with a dynamic problem. That is a work order, not a statistic.

2. Benchmark against yourself over time, with an index. Report RevPAB as an index against the same period last year rather than as an absolute:

RevPAB index = RevPAB this period / RevPAB same period last year

This handles seasonality, which matters enormously in PBSA and student housing where the September cycle dominates, and it strips out the definitional arguments because both figures use your definitions.

If you do want to compare against an external figure, ask five questions first: is it net of utilities, does the denominator include out-of-service beds, does it include ancillary revenue, is it bed-nights or bed-months, and what is the average stay length. If you cannot get all five answers, the comparison is decorative.

The reporting cut that actually gets used

Report RevPAB, RevPOB and occupancy together, at these levels:

• Bed: for outliers and structural problems
• Room type: for pricing decisions
• Property: for site management accountability
• Cohort: for like-for-like comparison
• Portfolio: for the board

Weekly at property level, monthly at cohort and portfolio level, with the bed-level outlier list reviewed in the same meeting that owns maintenance and pricing. A metric reviewed in a meeting with no authority to act is a metric that will not move.

For the wider set of bed-level measures that sit alongside yield, see our post on per-bed reporting KPIs, and on bed-level pricing strategies for the rate side of the identity. Definitions for RevPAB, per-bed inventory and void period are in the shared living glossary.

Calculating any of this requires bed-level inventory in the first place, which is the point our inventory and occupancy and reporting and analytics modules are built around.