Reading Demand From Tour Data: What Aggregated Tour Patterns Do and Do Not Tell You

The national demand series are rigorous, public and well documented. They are also national — and the national average is exactly where the signal an operator needs stops existing.

What the published demand data actually covers

Start with the best-documented part of this business. The Census Bureau and HUD New Residential Construction release for July 2026, published August 18, put completions in buildings with five units or more at 329,000 at a seasonally adjusted annual rate, down 14.8 percent from June’s 386,000 and 25.6 percent from 442,000 in July 2025.

Neither decline is statistically significant at 90 percent confidence. The month-over-month interval is plus or minus 28.0 points, the year-over-year interval plus or minus 27.2, and both straddle zero. That is the release telling you, in its own published terms, that a single month of this series cannot carry a narrative.

Census Bureau and HUD, New Residential Construction, seasonally adjusted annual rate, 2026; July is preliminary and May and June are revised; month-to-month moves in this series are noisy and the July decline is not statistically significant at 90 percent confidence; the cleaner read is the not-seasonally-adjusted year to date, 242,100 units in January through July 2026 against 280,100 a year earlier, down 13.6 percent.

The cleaner read is cumulative. Not seasonally adjusted, 242,100 units were completed in five-plus-unit buildings from January through July 2026 against 280,100 a year earlier — down 13.6 percent, interval plus or minus 12.1. That difference is significant. Annual not-seasonally-adjusted completions ran 591,700 units in 2024 and 467,700 in 2025. In the same release, starts came in at 421,000 SAAR, permits at 490,000, and units under construction at 666,000, down 4.4 percent year over year.

One measurement note matters more than it sounds: this series counts housing units completed, not units leased or occupied. It is not absorption. Treating a completions figure as a demand figure is the most common misreading of this release.

For absorption you need a different source. RealPage Market Analytics data showed net absorption above 187,000 units in the second quarter of 2026 and annual demand of roughly 271,300 units for the year ending that quarter, below the decade average of about 340,000 and dragged down by net move-outs in late 2025. Trailing-year supply came to roughly 340,200 units, a sixth consecutive quarter of declining annual supply after deliveries peaked near 588,000 in late 2024. Occupancy stood at 95.5 percent, up a second straight quarter but down 20 basis points year over year, the South the only region below 95 percent. RealPage publishes this from its own tracked market-rate stock; it is not a census.

Census measures vacancy separately. The Housing Vacancy Survey for the second quarter of 2026 put the national rental vacancy rate at 7.3 percent, not seasonally adjusted, interval 7.1 to 7.5. The first quarter was also 7.3; the second quarter of 2025 was 7.0, and that difference is not significant. By region: South 9.5 percent, Midwest 6.9, Northeast 5.9, West 5.3, principal cities 8.0 against 6.9 in the suburbs. The caveat that matters: this covers all rental housing, including single-family rentals and small buildings, so it is not comparable to an operator occupancy report.

Why the national number is the least useful number you have

Here is the teaching point of this piece, and the concession data makes it unusually clean. In the Zillow July 2026 Rent Report, 39.8 percent of rentals on Zillow offered a concession — up 0.1 percentage point month over month and up 3.8 percentage points year over year from 35.9 percent in July 2025.

Now open the number up. Charlotte was 68.1 percent. Buffalo was 8.2. Both sit inside the same 39.8. Denver was 67.2, Salt Lake City 66.5, Dallas 65.6, Raleigh 65.4, Austin 65.1; at the other end, Providence 11.4, New York 17.5, Virginia Beach 18.9, Milwaukee 19.7. The largest year-over-year increases were Las Vegas up 15.2 points, Cincinnati up 13.7, Birmingham up 10.3 and Salt Lake City up 10.1.

Zillow Research, July 2026 Rent Report; measures the share of active Zillow rental listings advertising any concession, not the size or value of it; reflects Zillow’s listing mix rather than a census of apartment communities.

An operator in Charlotte who reads 39.8 percent and concludes the market is normal has been misled by an average. One in Buffalo who reads the same figure and concludes they are behind on concessions has been misled the other way. The national number is not wrong; it answers a question about the country, and no one operates a lease-up in the country. Two caveats belong on that chart: it measures the share of listings advertising any concession, not the size or value of one, and it reflects Zillow’s listing mix rather than a census of apartment communities. Zillow publishes no average concession size, so a claim that the typical offer is one month free is not sourced here.

Rent growth makes the same point in a sharper form, because the two best national series disagree. Apartment List’s National Rent Report for August 2026 put the national median at $1,390, up 0.1 percent month over month for a seventh straight increase and down 0.8 percent year over year — up from a record low of negative 1.6 percent in April, and the first positive August reading since 2022. Zillow, in the same window, had its Observed Rent Index at $1,962, up 2.3 percent year over year, the fastest annual pace in over a year, with multifamily specifically at $1,786 and up 1.7 percent and single-family at $2,314, up 3.0 percent. Zillow forecasts multifamily rents up 1.9 percent for full-year 2026.

Two more Apartment List readings are worth carrying forward. Its vacancy index sat at 7.1 percent in August 2026, down from a 7.3 percent peak in February, the first decline since late 2021 — but it covers stabilized properties only, six or more months on the platform and having reached 85 percent occupancy at least once, a different universe again from the Census 7.3. Time on market, list to lease, ran 32 days, up two days month over month and three year over year, the longest of any August since tracking began in 2019.

The benchmark that does not exist

Every series named above has a published methodology, a stated period, a defined universe and, in the Census cases, a confidence interval. Now consider the numbers operators are handed most often: lead-to-tour conversion, tour-to-lease conversion, no-show rates. There is no methodologically transparent, publicly documented national benchmark for any of the three.

Trace any figure in circulation and the trail ends at a vendor blog or a conversion calculator. No published sample. No stated period. No definition of what counts as a tour — whether a missed appointment counts, whether a walk-in counts, whether a repeat visit by the same household counts once or twice. Those choices move a conversion rate on their own, which is why a number without them is not a benchmark. It is a marketing artifact wearing a decimal point.

The same applies to two claims repeated so often they read as fact: that self-guided touring raises lease conversion by some specific percentage, and that a specific share of tours happen after office hours. Neither traces to a primary document, and both are correlation presented as causation. A prospect who books and completes a tour independently may simply be further along already — the format may be sorting for readiness rather than creating it. You cannot tell which from a percentage with no study behind it, so this article prints no conversion figure.

The constructive alternative sits in your own portfolio. Aggregated and anonymized across many tours, your tour data read against your own trailing twelve months, in your submarket, with your seasonality, is a real baseline. It has a defined universe, a known period and a consistent definition of a tour, because you set all three. Nothing about it describes an individual prospect, and nothing about it needs to. A borrowed number has none of those properties. Where sensing is involved, presence sensors are the source of the counts, and availability is built state by state.

Reading your own aggregated tour data

On step three, the seasonality evidence is thinner than the confidence around it. Zillow reported in May 2025 that website traffic, messages to landlords and applications on Zillow Rentals typically peak in the first week of June, with listing views, applications and outreach to property managers all peaking there in each of the prior two years. That measures platform search traffic, not property tours, and it is 2025 data. Apartment List, meanwhile, described August 2026 as the tail end of peak moving season. Useful context. Not a curve to staff against.

None of this makes the national series unworthy of attention. A 13.6 percent decline in year-to-date completions, a sixth straight quarter of falling annual supply, occupancy grinding up two quarters running, concession share rising 3.8 points: together those describe a market where new supply is thinning while incentives keep widening. That tells you what kind of year is underway. It does not tell you what your Tuesday tour count should be, and no published number does. That is the work your own data has to do — roughly how we at Tourzy think about aggregated tour reporting.

Frequently asked questions

Is there a published national benchmark for lead-to-tour or tour-to-lease conversion?

No. The figures in circulation trace back to vendor blogs and conversion calculators that do not publish a sample size, a time period, or a definition of what counts as a tour. Without those three things a number cannot be checked or reproduced, which means it cannot be used as a benchmark. The honest substitute is your own trailing twelve months of aggregated, anonymized tour counts in your own submarket, which is a baseline you can actually verify.

Why do Apartment List and Zillow report different national rent growth?

They measure different things. Apartment List reported a national median rent of 1,390 dollars in August 2026, down 0.8 percent year over year, using a same-unit repeat-transaction model built on transacted rents. Zillow reported an observed rent index of 1,962 dollars in July 2026, up 2.3 percent year over year, built on asking rents in active listings across all rental types. That is a definitional gap, not a contradiction, and the two numbers should not be averaged together.

Can the Census rental vacancy rate be compared to my occupancy report?

Not directly. The Census Housing Vacancy Survey put the national rental vacancy rate at 7.3 percent in the second quarter of 2026, but that figure covers all rental housing, including single-family rentals and small buildings, not just apartment communities. RealPage Market Analytics reported occupancy of 95.5 percent for the same quarter across its tracked market-rate stock. The two are measuring different universes, so treat them as separate readings rather than as one number that disagrees with itself.

What does aggregated tour data actually tell an operator?

Read in aggregate and anonymized across many tours, it tells you the shape and direction of demand at your own communities over time. It shows whether tour volume this month sits above or below the same month last year, whether a floor plan draws steady interest or thin interest, and whether a seasonal turn is arriving early or late. It does not describe any individual, and it is not a substitute for a leasing conversation.

Is national apartment supply still rising in 2026?

The published series point down, though with caveats. Census and HUD counted 242,100 units completed in buildings with five or more units from January through July 2026 on a not-seasonally-adjusted basis, against 280,100 in the same months of 2025, a decline of 13.6 percent that is statistically significant. RealPage Market Analytics reported roughly 340,200 units delivered in the year ending in the second quarter of 2026, a sixth consecutive quarter of declining annual supply.

Sources: U.S. Census Bureau and HUD, New Residential Construction, July 2026 · U.S. Census Bureau, Housing Vacancy Survey, Q2 2026 · RealPage Market Analytics, 2Q 2026 Data Update · Apartment List National Rent Report, August 2026 · Zillow July 2026 Rent Report · Zillow, rental search seasonality, May 2025