The KieranTimberlake office is in a renovated mid-20th-century industrial building in Northern Liberties, an old neighborhood of central Philadelphia. Two stories tall, the building originally housed a bottling plant for a brewery, and the upper floor—now the studio—is a vast open space lit by strip windows and a roof monitor. When I first saw it in January 2015, the firm had just moved in. At the time, Stephen Kieran, FAIA, referred to the new office as “our huge sandbox.” Some of the toys included mobile workstations, adaptable meeting rooms, and plenty of flex space. The most radical innovation, in the service of energy conservation and green design, was the cooling system. Instead of air-conditioning, there was a combination of passive and active techniques: heavy insulation, automated sunshades, operable windows, natural and mechanically assisted ventilation, and dehumidification. Because it was winter when I visited, it was impossible to know how well it would all work. “This is a daring experiment in Philadelphia, where summers are hot and humid,” I wrote at the time. Three cooling seasons later it’s time to ask: How did the experiment fare?

Air-conditioning has been a part of American life since 1902, when Willis Haviland Carrier, a recent Cornell engineering grad working for a firm in Buffalo, N.Y., invented a mechanical dehumidification and air purification system for a Brooklyn printing company. Interestingly, what drove his invention was not the workers’ comfort but the integrity of the industrial process—high humidity was causing the printing company’s paper to swell. In fact, almost all of the early air-conditioning installations were for bakeries, candy factories, textile mills, and cigar and cigarette plants, all processes that benefit from humidity control. Air-conditioning for human comfort soon followed. As early as 1903, Alfred Wolff, who had earlier installed an ice-based ventilation system in New York City’s Carnegie Hall, used ammonia-absorption machines to dehumidify the trading floor of George B. Post’s newly constructed New York Stock Exchange building. Other early installations included movie theaters. In 1928, an air-conditioning system—designed by Willis Carrier—was installed for the first time in a high-rise office building.

In a typical office building, lighting, heating, and cooling represent as much as three-quarters of total energy consumption (depending on the local climate), so eliminating air-conditioning has potentially massive ramifications. The KieranTimberlake (KT) office was clearly prepared for the trial run. During the building renovation, which took two years, several hundred thermal sensors were embedded in the fabric of the building to enable the monitoring of temperatures, both on the surface and within the floors, walls, and roof. Dissatisfied with available off-the-shelf technology, the firm developed its own miniature wireless sensors and produced them in-house. (The sensors are now part of a KT product called Pointelist, a wireless network that monitors the thermal performance of workplaces; kits are currently being beta-tested.)

Measuring temperature and humidity is only half of the thermal comfort equation; the other half is finding out how people actually feel. To that end, the 120 KT staff were sent an online survey three times a day from April to October 2015. The questionnaire asked people to rate their current condition (“much too cold,” “warm but comfortable,” “too hot,” and so on), describe what they were wearing, and indicate where they were physically located in the building. Instant feedback was provided so that individuals knew where they stood compared to the entire office. (The survey has resulted in another product, called Roast, a cloud-based app that provides building managers with a way to ascertain occupants’ comfort levels.)

There was space on the questionnaire for comments, which produced some lively feedback ranging from the cheerful “I’m wearing my fanciest Hawaiian shirt today in honor of the warm weather,” to the slightly cranky “Ask not what the temperature survey can do for you; ask what you can do for your temperature survey.” As spring turned to summer and outdoor temperatures rose, however, the tenor of the remarks changed: “Large fan to the left of me, small desk fan to the right of me, sweat on the back of me,” and the succinct “I am physically melting.” The atmosphere in the studio was graphically described by this plaintive comment: “It’s like being in the Amazon rainforest but with phone noises in place of animal sounds.”

An Experiment Gone Awry
What was going wrong? The cooling system was designed to draw cooler night air into the building, dehumidify it, and then circulate it through a plenum beneath the raised floor of the studio. It wasn’t working. The concrete floor slab beneath the plenum was intended to remain cool, so that in the daytime cooled air could be supplied via individually controlled floor registers, but it turned out that the thermal transfer between the air and the concrete was less effective than anticipated. In addition, the overnight period wasn’t long enough for there to be adequate dehumidification. The firm tried a different strategy: “flushing out” the studio space directly. At the end of the workday, warm air was evacuated via the overhead monitor, and cool night air was brought in through open windows. The windows were closed in the morning, keeping the cool air inside. This worked better, but as the summer nights became warmer, flushing was less effective and indoor temperatures increased yet again. The number of “much too hot” survey ratings increased, too.

At the time, the firm was working on a major project, the new American embassy in London. “We had a meeting scheduled with representatives of the State Department,” Kieran says. “The outside temperature was 97 degrees, so we suggested that they wear shorts, which they did. We managed to bring down the interior temperature a couple of degrees, but it was still pretty hot.” The image of visiting diplomats in short pants calls to mind a scene from a P.G. Wodehouse novel.

Various measures were taken to mitigate the heat. Large floor fans were placed in the studio and individual desk fans were distributed to the staff. The dress code was loosened to encourage lighter clothing. Flex work times allowed employees to come in early when the building was cooler. On selected days, water ice—a South Philadelphia summer staple—was distributed free. For a significant number of people, however, it was not enough. The survey responses showed that while some tolerated the heat, others felt uncomfortably warm even at relatively low temperatures. “Human beings turn out to have varying responses to heat and cold,” Kieran says. “We’re all really different. That’s the problem with averaging out.”

The Advent of the Cool Room
“Averaging out” originated in the 1960s when the American Society of Heating, Refrigerating, and Air-Conditioning Engineers (ASHRAE) established thermal guidelines for air-conditioned office interiors. The guidelines were based on what 80 percent of occupants felt was comfortable; the prototypical subject was a 154-pound man in a business suit—think Mad Men’s Don Draper. These guidelines, with periodic modifications, have lasted more than five decades, but in 2015 a pair of Dutch researchers published a paper on the gender aspects of office comfort that called the ASHRAE premise into question. The researchers tested the metabolic rate for women engaged in office work and found that the rate was lower than the industry standard, and that women consequently tended to find conventional cooling temperatures uncomfortably low.

Forty percent of the KT staff is female, and I asked Roderick Bates, a member of the in-house research group, if his survey supported the Dutch finding. He responded that the break point temperature at which 80 percent of the staff indicated that they still felt comfortable turned out to be 83 degrees—considerably warmer than the industry standard of 68 to 78 degrees—but there was no indication of gender differences. “What we’ve found so far is that women are much more likely to participate in the comfort survey but they don’t differ greatly in comfort response for a given temperature relative to men,” he wrote in an email.

“I was expecting a significant difference and I wonder if it may be due to our more accommodating dress code that finds men and women wearing clothing of similar thermal values whereas in the past men wore outfits with a higher thermal value.”

Even at 83 degrees, the 80 percent acceptance rate still left a couple of dozen people in the KT office feeling uncomfortable. To accommodate this sweaty group, an improvised air-conditioned workspace was created on the first floor and kept between 72 and 74 degrees. Some used the so-called Cool Room intermittently, while others moved in for the duration of the summer; according to the survey, about 20 people regularly worked there. Unexpectedly, despite their lower metabolic rates, women were disproportionately over-represented. The explanation? Gender is only one variable in how people experience thermal comfort—age, weight, and body type probably also matter, as well as cultural factors such as upbringing and where you grew up (KT staff members come from all parts of the country: from the hot South to the temperate Pacific Northwest).

At the end of that first summer, everyone felt a sense of relief. There were very few complaints during the ensuing heating season, even though the office was kept at a relatively cool 73 degrees. “There is no such thing as bad weather, only wrong clothes,” goes an old Scandinavian saying, but of course it refers to cold weather. “Cold is much easier to deal with than heat,” observed Bates. “Even though the winter inside/outside temperature differential is larger (40–50 degrees) compared to the summer (20–30 degrees), people deal with cold easily by simply adding a layer of clothing.”

The Human Problem
Meanwhile, the partners at KT had to face a hard truth—their cooling experiment had failed. The solution was obvious: The following spring they installed a cooling tower and air-conditioned the second floor. The ducted ventilation system had been designed to accommodate air-conditioning, and the raised floor in the studio meant that air could be circulated without any modifications to the workspace itself. “What swayed us to make the decision was the data we compiled, especially the thousands of staff surveys,” Kieran says. “There were simply too many people who were not at all comfortable during too many hours of the summer. Some health issues arose. The heat generated by people and equipment, especially the computers, proved more difficult to overcome than we had anticipated.” (A human body gives off about as much heat as a 100-watt lightbulb; an average desktop computer can give off 75 watts, while a large monitor can add up to 500 watts.)

Air-conditioning commenced in July 2016. The survey responses reflected a generally upbeat reaction. “So pleasant in here. I don’t even need my usual vat of iced tea this morning,” “The best thing about cooling is that there is no fan noise,” and “A girl has no sweat.” Some people missed the open windows, and there were still a few outliers who felt uncomfortably warm, but on the whole the studio was a happier place.

Kieran says that the decision to resort to air-conditioning was not an easy one. “We really had high hopes that we could make the building work without air-conditioning. We thought we had a good chance because the original bottling plant was designed for natural ventilation with operable windows and venting through the ridge dormer,” he said. “We tried, we failed, and we learned.”

What they learned is how to operate a dual-mode cooling system. As long as outside nighttime temperatures are below 67 degrees, the interior can be flushed out and cool nighttime air circulated through the plenum, enabling the building to function entirely with natural ventilation. Warmer weather requires air-conditioning, although the set point is 81 degrees, well above the industry standard. This temperature represents what 80 percent of the staff reported as “comfortable,” and a permanent Cool Room continues to offer an option for those who fall outside the norm. The rest of the first floor, which houses the workshop and offices, is not air-conditioned. “It’s generally 4 to 5 degrees warmer than upstairs,” Bates wrote me. “This last summer, large floor fans were deployed and comfort requirements were met for the most part. If things got too warm, people typically went upstairs to work for a stretch.”

I asked Bates how effective the dual-mode system has been in reducing energy usage. “We’re in the process of creating a model to allow just such a calculation. However, we have projections based on our analysis of the comfort survey data that found that during the 2015 summer interior temperatures exceeded the comfort threshold for 80 percent of the staff for approximately 25 percent of the work hours,” he said. “From this we are assuming that we can meet the comfort needs with passive means 75 percent of the time, running the AC only for the remaining 25 percent of the hours, which is a significant energy reduction.”

“The real question is whether we can un-adjust from air-conditioning,” Kieran says. That un-adjustment involves learning a variety of new habits: how to live with indoor temperature swings, changing clothing seasonally, and using natural ventilation—that is, opening windows. It sounds simple in theory, yet as Kieran observes, “Behavioral adaptation proved to be the most difficult part of our experiment.”

He explains that clothing and work-schedule changes were difficult to implement. Some of the staff were resistant to the idea of “dressing down”—shorts and T-shirts are not to everyone’s taste. Likewise, flexible work schedules that were adapted to temperature swings did not suit everyone. Once people experienced open windows, they complained when windows were closed after a nighttime flush-out. Others found street sounds distracting. The noise of fans was universally disliked. Nor was the idea of a Cool Room universally popular—some referred to it as the FEMA Room. For those who were accustomed to the ice-box-cool interiors of American offices and shopping malls, “warm but comfortable” felt like too much of a compromise.

That is a key lesson of the KT experiment. The most challenging aspect of energy conservation may be not the technological devices but ourselves. Changing human behavior is never easy, whether it involves consuming less, driving less, living smaller and leaner, or working in an office that’s not quite as cool.