I’ve been in CRE for just shy of 20 years. I’ve been a consultant for half that time, receiving client data and trying my best to parse it so that we could help them make decisions. I’ve been on the occupier side for the other half of that time, managing data and doing my best to parse it so that we could make good decisions. Throughout that time, there was never a moment when the data was ready to simply pick up and use without any manipulation.
This is not a knock on the company I was with or the clients I worked for. It’s a statement about how real estate works versus other functions. In finance, the databases keep auditable, detailed journal entries of every activity that takes place. A person can sit down with that data and piece together innumerable realities about what took place and how current circumstances came about. The data itself is capable of telling the story and therefore AI can be trained to read that story.
In real estate, the data typically is a description of physical space, the people who occupy (or may occupy) that physical space, and some of the activities that took place in that space. Real estate data is like archeological records. They provide evidence of things that happened in the past, but there are often multiple interpretations of the information. A good archeologist will have studied past trends to understand the most likely paths that led to the existence of the information they have found.
Good real estate analysts must do the same. Let us look at a hypothetical example to see why this is so important.
The scenario: There is a 10,000 square foot office in the US with 67 desks (149 square feet per desk). There are 100 employees assigned to this office. Over the past year, peak occupancy has been 50 employees on both Wednesdays and Thursdays. Peak occupancy has occurred 1 time every month. Average occupancy is 25 employees across all days with the lowest occupancy on Mondays and Fridays. The lease for this site expires in 18 months, and the team needs to figure out how much space they need for a future office.
There is no gotcha built into this hypothetical. This exact scenario probably feels familiar to most real estate people in the US.
What typically happens (assuming a data-driven decision and not a political one): There’s a peak of 50, so let’s throw 10% on top of that to be safe. We therefore need 55 desks which will require about 8,250 square feet at our current 150 square feet per desk standard. Going to market, the broker looks for any spaces from 8,000 square feet up to 10,000 square feet (you couldn’t possibly fit smaller, but anyone can go bigger, of course). You find a great space with financial savings as 9,250 square feet and you can conveniently fit 62 desks into the larger space. The project ends up breaking even versus the past budget saving the cost of inflation if you had done a renewal. It feels like a win-win.
What can happen with a little investigation: It turns out badge data was used to measure the 50 employee peak occupancy. This was overstating the number of employees using desks on a peak day by 15%. The actual peak desk usage was only 43 employees. Because peaks do not occur regularly, you can plan for 10 desks to be smaller because they are only likely to get used a couple of times a month. This brings your square foot per desk down from 150 square feet to 135 square feet without sacrificing any amenity spaces or working space for those employees who regularly use the office. Including a 10% buffer on peak occupancy, 47 desks are needed at the new 135 square foot standard yielding 6,345 square feet required. Going to market, the broker looks for spaces from 6,345 square feet and up (you couldn’t possibly fit smaller) and finds you a 7,250 square foot space. Because you have confidence in your desk occupancy numbers, the additional square footage goes towards additional collaboration, meeting, and amenity space making the office more desirable. Because of the 28% reduction in square footage, the project ends up saving money versus the previous budget while delivering an office that feels more vibrant with better amenities.
This is not to knock the first scenario. That team still leveraged their data to create a beneficial outcome to the business by better aligning the space to the data. In the current world, that is still hitting a double (to use a baseball metaphor to imply a really good outcome). But by simply having an increased grasp on the data, so much more can be done. But that other data does not come from a system. It often comes from additional research, follow-up questions of the local teams, better understanding of how space could be used across times, and having leadership willing to allow creative solutions that do not simply treat all spaces as if they must be the same.
Good data managers in real estate do the basics of any good data manager: keep their data up to date, validate and test data quality, provide value-added reports to key stakeholders, and maintain good data standards. However, the best ones are also constantly seeking to understand the data they do not have access to. They try to understand the impact of human behavior on the data they have.
At the end of the day, real estate is all about people. Our data is a reflection of the people we support in our physical spaces. When we do the right things, those people get better outcomes.
***
Editor’s note: Somehow this this the first post in all my years where I have felt the need to add the category of ‘data management.’ I suspect I may need to revisit this topic a few times in the near future…