Spacing needs are not as straightforward as they once were. To address the growing complexity of organizations' space optimization, occupancy analytics is an emerging segment.
Before the COVID-19 pandemic, most employees commuted to work each day, and office buildings were bustling with people. Then, suddenly, almost everyone was forced to work from home for more than a year, and commercial buildings were left empty.
Now, more than three years after the onset of the pandemic, even though offices are open just as they were before 2020, many people still work from home.
As a result, organizations' office needs have changed. Figuring out those needs, however, is complicated. Some people are in the office every day, others work from home every day, and still others work from home on some days and in the office on others.
Many enterprises are, as a result, turning to occupancy analytics to determine their spacing needs.
Even before the pandemic, however, space optimization was an emerging concern. Technology had reached the point at which IoT sensors and Wi-Fi could enable organizations to track movement patterns.
Colleges and universities, therefore, began using occupancy analytics to figure out how libraries and other common spaces were utilized. Museums tracked movement to know which exhibits attracted the most visitors. Commercial real estate firms analyzed which stores attracted the most foot traffic. And any type of organization could use movement and occupancy patterns to minimize energy consumption.
To meet the organizations' space optimization needs, a group of vendors specializing in occupancy analytics has emerged. Among them are companies such as Colliers, Density.io, Lambent and Spaceti.
Recently, Richard Scannell, CEO of Lambent, discussed occupancy analytics, including its brief history, growing importance and long-term outlook. In addition, he spoke about different applications of occupancy analytics and the data used to inform space utilization insights.
Editor's note: The following has been edited for length and clarity.
What is the definition of occupancy analytics?
Richard Scannell: It's trying to help people understand how their spaces are being used. There's a subtle difference between tracking people in spaces and tracking spaces and how they're used by people, and occupancy analytics is the latter. This segment is not in the business of tracking whether a given person is in a building but how the building is being used by all the people who are using it -- or not using it, as the case may be.
Within that, there are two different words with specific meaning. There is occupancy, which is the number of people in a space. Then there's utilization, which is reflected by the context of a space. If there are 50 people in a building, that's important -- it's occupancy. If that building was designed for 60, that's good utilization. But if it was designed for 500 people, that's poor utilization. Occupancy is the raw count, and utilization is the count in the context of the intended use of that space. Occupancy analytics examines utilization. It's an MRI for space.
What is the purpose of occupancy analytics?
Scannell: The obvious things are looking at whether people are coming to the office, whether space is being used, which space is most used, what are the busiest hours of the day and week, the busiest month of the year, what happens after a snowstorm and so forth. But that all has a limited shelf life as we try to figure out what this return to work looks.
However, once you have all this data, you start to realize other things you can do with it.
What's an example of other things a user can do with occupancy and utilization data?
Scannell: We have a client that's a big university and has a big library. Within the library are tables with three seats on one side and three on the other. They observed that when a student sat in the middle seat on one side, no one else would join the table. It had a land-grab effect, and other students looked for other spots.
So they put big plants across the middle of the table to create an artificial break. Suddenly, there were more students sitting at the table, which they were able to observe using occupancy analytics. They saw the occupancy of that set of tables and how it went from one number to another. That's just a simple example.
The long-term look is about gaining control. If you want to have a smart building, you have to know when people are there and where they are. You can build all the flexibility you want into a building control system -- HVAC, power, water -- but if it doesn't understand when and where people are present, it's [not useful]. Presence is an important thing to understand. That's where occupancy analytics can have a sustainability impact. There are all kinds of ways this information can be used from near-term, obvious stuff to this concept of a living, smart building.
When did the concept of occupancy analytics emerge? Is it something that resulted from the COVID-19 pandemic when more people started working from home -- and haven't returned to the office -- or did it emerge before that?
Scannell: There are different segments within occupancy analytics. There are folks that are sensor-driven and then others like us that use Wi-Fi to track occupancy. With IoT, which began before the pandemic, the concept of occupancy became a hot topicand some of the companies that use sensors to track occupancy started to take off and get unicorn valuations.
What's happened since then is, customers have begun to think about their portfolios in a stratified way.
Scannell: We think of the real estate portfolio of an average company as a pyramid. At the top may be spaces that you don't want anyone to track, such as the CEO's office in a Fortune 500 company. The next set of spaces are those you care about that may be expensive and sensitive in nature, and there are reasons you want to know exactly what is happening [with IoT sensors]. At the bottom are spaces such as a bank branch where if someone walks in the door, you know they're occupied.
In the middle is the thick part where you want to know more than you know today, which is almost nothing, and you want to do things like make a building smarter. But the cost of retrofitting that space with IoT sensors is prohibitive. If you deploy sensors, you're going to get much more accurate data than using Wi-Fi or other methods.
But with Wi-Fi, you'll get about 85% accuracy in terms of how many people and where they are. That's good enough when there might be 500 people in a building or 520. If it's a vault, 85% accuracy is not good enough. There are spaces where sensors make a ton of sense, and there are spaces where good enough is good enough. Now it's about fit for purpose.
Richard ScannellCEO, Lambent
Why did occupancy analytics emerge? Was it simply a matter of finally having the technology such as IoT sensors and ubiquitous Wi-Fi?
Scannell: I think so. For 100 years, if you gave me your office application model, such as a VP getting a [400-square-foot] corner office with plush carpeting, an engineer getting a [100-square-foot] cubicle, and so and so forth, I could tell you how much space you need. It was a linear calculation. Everyone came to work on Monday morning, and the week ended on Friday evening, and that was it.
Now there are three factors that have changed things. There were ubiquitous Wi-Fi and nearly free video conferencing that were the technology catalysts, and then the pandemic pushed things over the edge. Now no corporate building is going to exist the way it used to, so you can't have a linear relationship between employees and space. Real estate people used to have a tried and trusted model that was predictable. Then one day they woke up and didn't know what the heck was going on. The question became how to get eyeballs out into these facilities.
Before IoT sensors and ubiquitous WiFi and the advent of occupancy analytics, how did organizations attempt to determine space utilization?
Scannell: You've probably seen those clickers that people hold. You would hire consultants who would sit at the door or walk around. It was spot-checking, but all it did was count occupancy at a point in time. It was bit like taking the national census. Directionally, it might give you some insight. But they could do that spot check during a bad snowstorm or on a sunny day in spring when everyone is out at the beach, so they were unreliable.
There are other proxies as well. The one everyone is familiar with is swiping a badge when you enter a building. But that doesn't tell how many people are in a building. You know how many people came into a building. But you don't know if they came in and turned around 15 minutes later and walked out. You don't know if the same person went in and out 10 times to get coffee or smoke a cigarette. You don't know where they went in the building. You don't know how the space in the building is being used -- did everyone go to the northeast corner of the second floor because that's where there's nice, new furniture?
What data drives occupancy analytics?
Scannell: We are tracking space, not people. Most of the products in our sector connect into the Wi-Fi network, typically through an existing data lake that is collecting data from the Wi-Fi network. We then pull all that data into our platform and see all the connected devices. Then comes the hard part -- the special sauce -- which is figuring out which devices represent people. We use AI to understand the movement of devices throughout space.
There are things like printers that never move but are connected to the network, so after 24 hours or so, you can assume that's not a person. Then there are devices [such as laptops and mobile phones] that have volatility that come in together but then maybe separate for a while, and one remains stationary while the other moves around, and then they come back together again and ultimately leave together. What we are looking for is the most human-like device, and then we count that device.
Can occupancy analytics be predictive and prescriptive, or is mostly descriptive?
Scannell: We have some thought process around something that is not quite predictive but is forecasting presence where we don't have data. We may not have data about a particular part of a building, but we can forecast what we believe is happening. We have not implemented this, but we have some thought around that.
What you can do, like any statistical model, is track patterns, such as what happens the day after a snowstorm of more than six inches, and make reasonable assumptions.
What are the advantages of specialized occupancy analytics platforms such as Lambent Spaces over general use analytics platforms or homegrown data science models?
Scannell: It's what we call disambiguation. It's the multiple devices connected to the network.
Occupancy analytics is a big data problem combined with an intelligence problem -- how you make sense of things that don't have linear relationships. The barrier to entry is time and effort, and we and our competitors are doing some interesting stuff. That's what makes a market. We have some customers that, at first, did some homegrown stuff out of desperation coming out of the pandemic, trying to get some sense of what was happening in their facilities around of the world. But by and large, they found [their homegrown effort] unsupportable and unscalable.
What is the long-term outlook for occupancy analytics? Is it something that addresses a major need coming out of the pandemic but will then become less important in a few years, or does it have staying power?
Scannell: One of the most talked about things right now is getting people to return to the office. That's not about mandates. It's about the employee experience and creating a desire to be in the office and then understanding that on a continuing basis. Then I go back to that concept of a smart building. To have a smart building, you have to know where the people are. I believe that in our kids' and grandkids' time, sustainability will become one of the leading indicators of business success.
You can't have a building in Arizona cooled to 70 degrees if there's no one in there. It's going to be unacceptable behavior. The same if you heat an empty building in Maine to 72 degrees in the wintertime. Being sure a building can respond to the reality of its users will be critical, and you have to figure out how to do that for 20 million square feet of portfolio. The future is having enough capacity for needs, not having enough capacity for population.
Eric Avidon is a senior news writer for TechTarget Editorial and a journalist with more than 25 years of experience. He covers analytics and data management.