Making Sense of the Coworking Data
August 26, 2017
Since the beginning of 2017, I've been studying the coworking trend to see what — if anything — it has to tell us about the future of work, the workplace, and ultimately, the office real estate industry. This line of enquiry has produced many insights about where the office industry might be headed, but before we ponder tomorrow, it's helpful to understand where we are today. And for this, I turn to the data.
Unfortunately, reliable coworking data is hard to come by. It's difficult enough as it is to find data for traditional commercial real estate — a mature, trillion dollar asset class. Coworking, still in its infancy, is barely tracked. Through a little bit of elbow grease, however, I've managed to cobble together some different data sources to separate the signal from the noise.
Domestic Coworking Locations
Let's begin with the number of coworking locations in the United States.
We don't have much historical data, but we do know that the country's first coworking location is generally considered to be Spiral Muse, which opened in San Francisco in 2005. Thus there was one coworking location in 2005. By 2007, various estimates suggest that another dozen or so pioneers had entered the fray, bringing the number of locations to fifteen. At this point, the coworking movement began to pick up steam. Between 2007 and 2010, early adopters opened another 250 locations. But beginning in 2010, the coworking trend went exponential. By the end of 2016, it's estimated there were 3,600 locations in the United States. That's a sustained, 60 percent compounded annual growth rate over six years.
Global Coworking Locations
The growth story played out similarly across the entire globe. The number of global locations is estimated to have increased from 600 in 2010 to more than 11,000 at the end of 2016. One step further, forecasters at Emergent Research project that the number of global locations will hit 26,000 by 2020.
Domestic Coworking Square Footage
If you're in commercial real estate, a more interesting statistic than the number of coworking locations would be square footage occupied by coworking. Unfortunately, no one is yet tracking this metric, but I'm sure it's only a matter of time. In the meantime, I have found two market reports by JLL and Cushman & Wakefield which report domestic coworking inventories of 27 million and 40 million square feet at the beginning of 2016 and 2017, respectively. Extrapolating these numbers forward, it's probably safe to assume that coworking occupies approximately 50 million square feet as of today (mid-2017).
Fifty million square feet does not sound all that significant in the scope of the four to ten billion square feet nation office inventory (the range depends on whether you include tertiary markets or owner-occupied and government office buildings). But do recall from the location data above that almost all coworking space has come online since 2010. During this same period, the entire United States office market has absorbed only 250 million square feet. This means that coworking accounts for approximately 20 percent of all absorption this cycle. That's significant. It's no surprise that WeWork has been the country's largest lessee for three years running.
As of mid-2017, coworking likely occupies around 50 million square feet of office, accounting for 20 percent of the nation's absorption since 2010.
Global Coworking Membership
Of course, locations and square footage quantify only one side of the market equation: supply. What about demand? In commercial real estate, the fundamental measure of demand is the vacancy rate, which is the percentage of existing square footage not leased by office tenants. In coworking, the business model is less focused on square footage. Instead, demand is measured by members.
I've not found domestic membership data, but researchers at deskmag track the global statistic. Worldwide coworking membership is estimated to have increased from 21,000 in 2010 to 835,000 by the end of 2016 — an 85 percent compounded annual growth rate over six years. It is believed that this number has already crossed the million-member milestone and will hit 1.2 million by year end. Forecasters at Emergent Research predict 3.8 million members by 2020.
Potential Domestic Market
Equipped with this knowledge, we can begin to consider the size of the potential domestic market from a top-down perspective. Let's assume (1) that the abovementioned forecast of 3.8 million global members is accurate, (2) that the US is approximately one-third of global demand, and (3) that the average square feet per coworking member is 100. This implies a potential domestic market of 130 million square feet.
Coincidentally, I've developed a bottoms-up forecast based on a segmentation of coworking demand into four groups — freelancers, remote workers, startups, and strategic enterprise users — which arrives at approximately the same number. Therefore, when people ask me how large I think the potential coworking market might become, my best estimate is somewhere between 100 and 200 million.
I estimate domestic coworking demand will end up between 100 to 200 million square feet.
Do note, however, that I'm using the word coworking here in the strictest sense of the term. For example, when IBM takes over an entire WeWork in lower Manhattan, I do not consider that to be coworking. At least not in the technical sense of the term which is supposed to refer to users sharing space. That being said, this IBM deal and other enterprise transactions like it are the very harbingers of the future of office I'm interested in.
We can also venture a guess as to when coworking in the United States might reach saturation. Remember that as of mid-2017, current coworking supply is likely to be around 50 million square feet. If you believe my potential market size above of 100 to 200 million square feet, that means that coworking inventory has room to double once or twice before supply exceeds demand.
Of course, this is a macro prediction whereas real estate is decidedly micro. Specific markets can and will saturate much sooner than the nation as a whole. Also, the history of real estate is nothing but one long series of booms and busts, caused by oversupply, so I expect empire-building coworking operators to overshoot the market before a correction rightsizes the industry. One final point: we humans tend to think linearly whereas the coworking trend is clearly exponential. Saturation could arrive much sooner than any of us expect.
Saturation could arrive much sooner than the gap between supply and demand would suggest.
I'll write about my segmentation of the users of (pure) coworking into freelancers, remote workers, startups, and enterprises.