Skip to content

Blog

The Spine of the Building Isn't the Elevator. It's the Data.

· Greg Gomer

The median office building that sold in the first quarter of 2026 was 39 years old. In 2000 it was 22. That's JLL data, by way of Propmodo's Franco Faraudo this week, and it's the single best explanation I've seen for why so many "smart building" projects stall out in year two.

Faraudo's piece is nominally about elevators. Nine million of the world's 23 million elevators are 15 to 20 years old, right at the age where Otis recommends refurbishment. More than a million of those are in the US and Canada. Owners carrying buildings through expensive debt and compressed valuations don't want a capital project, so they're doing partial fixes. A controller here, a drive there, buy five to seven years and hope.

Fine. Elevators aren't my beat. But buried in that piece is the best story about tenant experience I've read all month.

WHAT'S HAPPENING

Otis's president for the Americas, Joseph Armas, calls the elevator "the spine of the building." His point is that owners are bolting on access control, occupancy analytics and building management systems, and the 2008 elevator controller can't talk to any of them. An occupancy platform that can't see elevator traffic is missing the most reliable signal in the building about how people actually move.

Then the example. A New York finance firm's employees kept coming back late from lunch. Everyone assumed it was a people problem. Otis pulled the traffic data: the elevators couldn't handle the noon rush. The fix wasn't a new elevator bank. It was a 20% lunch discount offered 45 minutes before the peak.

Read that again. Behavioral data from a mechanical system, turned into a food and beverage promotion, to change when an entire firm rides the elevator. That is the whole REX thesis in one anecdote, and it came from an elevator company.

Here's what our buildings are showing us the same way. Across our US portfolio this fall, weekday business-hours door entries climbed about 17% from an early-October low to a late-October peak. After-hours entries did the opposite. Their share of all entries slipped from roughly 22% in mid-September to under 19% at the peak. More people, compressed into the same hours. That's a lobby problem, an elevator problem and a lunch problem, and you can see it coming three weeks out if you're looking.

The bookings tell the second half. In the week of November 8, roof deck bookings across the portfolio dropped to single digits after running around 100 a week through October. Outdoor space went from dozens a week to five. Nobody was surprised. It got cold. But the same week, lounge bookings hit their highest level of the fall, up about 40% from early October. Game room bookings ran at double their September-October average for the second week straight. Private office bookings roughly doubled. Large conference room bookings fell about 15% from mid-October.

The building reshuffled itself indoors in two weeks. The tenants didn't file a request. They just moved.

Leesman sees the same thing from the employee side. Nearly a third of office workers report no fixed attendance routine at all. Midweek is the peak, Mondays and Fridays are quiet, and 49% of people stay parked in a single setting all day once they're in. Rhythms, our Leesman research from last November, framed the workforce as an orchestra: everyone playing a different tempo. Your building is the hall. Most halls have no idea what's being played.

WHY IT MATTERS

The age number is the tell. A 39-year-old building has 39 years of systems that were never meant to share data. Elevator, access, HVAC, booking, food. Each one knows something true about your tenants. None of them knows it together. So the owner finds out about the noon crush when the renewal conversation opens with "our people hate the elevators." By then it's not an operations issue. It's a pricing issue.

Here's where I'd push back on the Propmodo piece. Faraudo frames modernization as a capital decision owners are deferring. It's a data decision owners are missing. The controller upgrade matters less than whether the signal from that controller reaches whoever runs the lobby, the café and the leasing tour.

WHAT TO DO

Treat every system in the building as a sensor first and a machine second. Access, elevator, booking, F&B. If one of them can't export what it sees, that's the upgrade to fund, before the cab finish.

Put the signals in one place and watch them weekly. A 17% rise in daytime entries with flat after-hours is a staffing and programming decision. A roof deck going dark while lounges and the game room spike is a reprogramming decision. You get two weeks of warning. Use them.

Steal the Otis move. When the data shows a peak, change behavior with an offer, not a memo. A lunch discount at 11:15 beats a sign in the elevator every time.

And stop calling it a smart building. The building was always smart. It's been telling you how people use it for 39 years. You just didn't have the plumbing to listen.

The spine was never the elevator. It's the data.