Multifamily KPIs: NOI per Door, Lease Spreads, Occupancy

Introduction

Two apartment portfolios can report identical new-lease and renewal-lease rent growth and still post very different same-store NOI, because the metric that quietly decides the outcome is resident retention. A property that keeps 65% of its expiring leases captures more of its rent roll at the higher renewal spread and avoids the make-ready cost, vacancy, and weaker new-lease pricing that every turned unit incurs. UDR made exactly this point its 2025 story: a turnover rate roughly 420 basis points below the prior year, the company told investors, was structurally responsible for a meaningful share of its same-store NOI outperformance versus the peer set. Reading a multifamily property well means reading the operating metrics as a system, not a list.

The institutional KPI set is small and consistently applied across public REITs, private platforms, and capital allocators: same-store NOI growth, NOI per door, the lease trade-out family (new, renewal, blended), occupancy and vacancy, and resident retention and turnover. Each captures one dimension of operating performance; together they show how a property is performing against underwriting and against peers. Trading-comp work uses them to rank REITs, M&A diligence uses them to judge a target's operating quality, and sponsor underwriting builds them straight into the pro forma that drives deal IRR. The table below anchors the definitions and where 2025 levels sat; the rest of the article works through how the metrics interact.

The cleanest single read on operating performance is same-store NOI growth, because it isolates organic property-level results from the noise of buying, selling, and developing. A total-portfolio NOI figure mixes in the effect of acquisitions and new development deliveries; the same-store cut strips those out by holding the property set constant.

Same-Store NOI Growth (Multifamily)

The year-over-year percentage change in Net Operating Income across the subset of properties owned in both the current and prior reporting period. By holding the property set constant, it measures organic operating performance and excludes the distortion from acquisitions, dispositions, and development deliveries that would otherwise inflate or depress a total-portfolio figure.

Where the boundary sits is the part analysts probe. A development delivery only enters the same-store pool after it has been stabilized and owned across both comparison periods, so a heavy developer can show muted same-store growth while its total NOI is climbing fast on lease-up. The 2025 same-store picture spread wide for this and other reasons: MAA sat near the bottom of the public peer range, pressured by Sun Belt supply absorption, while UDR's geographically diversified book kept it nearer the top. When you compare two REITs on same-store NOI, confirm the same-store definitions match before reading anything into the gap. The underlying NOI itself follows the standard build covered in how NOI and NCF roll up from a property's cash flows, so a same-store comparison is only as clean as the NOI definitions underneath it.

NOI per door strips portfolio size out of the comparison so two assets of very different scale can be judged on operating economics rather than headcount. The math is deliberately simple:

The figure varies far more by geography and asset class than by operator skill. Coastal gateway Class A in markets like Manhattan, San Francisco, and Boston runs $15,000 to $25,000+ per door; Sun Belt Class A garden-style runs roughly $8,000 to $13,000; Class B value-add product sits toward the lower end of those bands. Those gaps are rent-level and cost-structure differences, not performance differences, which is exactly why the metric is a normalizer and not a scoreboard.

Its real use is comparison. An investor weighing a 500-unit coastal Class A asset at $16,000 per door ($8M total NOI) against a 1,200-unit Sun Belt Class A asset at $9,500 per door ($11.4M total NOI) can set the headline NOI difference aside and focus on the per-door economics, the rent trajectory, and the cost base. Tracked over time, the same figure flags operating leverage: a portfolio whose NOI per door is outrunning the sector is converting rent growth into NOI faster than peers, while a flat or falling per-door figure signals structural underperformance no matter how large the portfolio.

Rent growth on the lease roll splits into two populations. New leases reprice units to incoming tenants at market; renewals reprice units for existing tenants who sign a new term. The blended trade-out is the weighted average of the two, weighted by how many leases fell into each bucket:

Blended Lease Trade-Out

The weighted-average rent change across both new leases (units leased to new tenants) and renewal leases (units re-leased to existing tenants), reported as a percentage and almost always on a same-store basis. It is the standard cross-REIT measure of rent-roll repricing, and the public disclosures break out the new and renewal components separately so readers can see which side is driving the blend.

Through 2025, renewals carried the blend. Existing tenants face real friction costs to move (deposits, moving expense, the time of an apartment search), so renewal spreads held up while new-lease spreads softened against new supply. Camden is the clearest illustration: by Q3 2025 its new-lease rates were down about 2.5% while renewals rose roughly 3.5%, netting to a blended figure near 0.6%. At the stronger end, American Homes 4 Rent posted Q2 2025 new growth around +4.1% and renewal growth near +4.4% for a blend of about +4.3%, with Essex around +3.0% and UDR around +2.8% blended same-store. The common pattern across the peer set was deceleration from the first half into the back half of the year, driven by Sun Belt absorption and steadier coastal conditions, with the better-positioned books landing in the +2% to +4% range. Because the components diverge this much, a single blended number tells you little until you see the new-versus-renewal split behind it.

Occupancy looks like the simplest metric and hides the most. Physical occupancy is just the share of units with a tenant in them. Most institutional portfolios run 94.5% to 96.5%, with best-in-class operators holding 96%+, and 2025 saw professionally managed stock tighten back toward the mid-90s as demand absorbed the supply wave. The non-obvious discipline is that operators do not try to maximize it. Pushing physical occupancy toward 99% almost always requires concessions and below-market asking rents, which leaves NOI lower than running at a deliberate 3% to 5% vacancy with rents held firm. The optimal point depends on submarket dynamics and where the asset sits in its rent band.

What ties occupancy to NOI is the gap between physical and economic occupancy, and that gap is where rent leakage hides.

Economic Occupancy

The percentage of gross potential rent the property actually collects, equal to collected rent divided by gross potential rent at market. It can sit several points below physical occupancy because of concessions, units leased below market (loss to lease), bad debt, and non-revenue units, which is why a property can be 96% physically full yet collect far less than 96% of its rent potential.

The wedge between the two is mostly loss to lease: the difference between what in-place leases charge and what the unit would command at today's market rent. In a rising-rent market loss to lease is positive and represents embedded upside that captures itself as leases roll to market; in a falling market it can invert. An analyst who quotes only physical occupancy misses the collected-revenue picture entirely, which is the distinction interviewers use to separate people who have read the metric from people who have only heard of it.

Turnover and retention are two sides of one number: retention is the share of expiring leases that renew, turnover is the share that vacate, and they sum to one. The reason retention sits at the top of the operating hierarchy is purely economic, not soft.

This is why operators who invest in resident services, amenities, communication, and competitive renewal pricing pull ahead of peers with identical physical product. UDR's roughly 420 basis point retention improvement in 2025 (and a turnover rate far below its ten-year average) is the cleanest public case study of retention discipline translating directly into same-store NOI outperformance, and it explains how the company posted the second-highest same-store NOI growth in the peer set despite a blended trade-out near the middle of the pack. The benchmark for the public REITs runs roughly 55% to 65% retention, with the best operators sustaining above that.

Read together, the metrics form a chain rather than a checklist. Retention decides how much of the roll captures the stronger renewal spread; the blended trade-out and occupancy feed the revenue line; that revenue, net of the operating cost base, becomes NOI; same-store NOI growth shows whether the property is improving organically; and NOI per door normalizes it against everything else. The same chain is what eventually flows up into the cash-earnings measures used to value the equity, which is why the REIT peer set that sets these multifamily benchmarks reports every one of these figures quarterly, and why NOI growth ultimately shows up in the FFO measure that captures REIT economics GAAP earnings understate. An analyst who can trace a single basis point of retention through to its NOI consequence is reading the property the way the operator does.