Recently, our co-founder and CEO Tushar Agarwal spoke on Bisnow’s panel regarding PropTech that’s disrupting the commerical real estate industry.
The main points addressed in this panel were:
- What drives decision making in real estate today? What is the role of the CTO/CPO/CDO in a modern property company ?
- How can technology support your portfolio, help foster better relationships, and boost returns treating tenants as customers?
- What is the future for the agent/broker?
- How can the use of data and technology help deliver a better end user experience?
- What are the challenges of dealing with data and can data be utilized for CRE opportunities and transactions?
Here’s a transcript of Tushar’s contribution to the panel.
Can you give us a brief overview of your business?
Tushar: I’m the founder and CEO of a business called Hubble at HubbleHQ.com. We launched in 2014 as an online marketplace to rent office space on a flexible basis and have been operating in London for the last three years. We help fill mostly flexible spaces, serviced offices, coworking spaces, excess capacity of people’s offices, as well as regeneration schemes across London.
What we’re moving into more now is building the world’s first artificial intelligence real estate company. Using the technology and data that we’re collecting, we’re creating smart insights that can start to automate and enhance the brokerage finance evaluation of: office space as of right now, and all of commercial real estate in the near future.
What are the main social media platforms that you use?
Tushar: Mostly, I use Instagram more than anything else now. I’ve stopped using Facebook and stopped using Twitter a long time ago as it seems to have become the cesspool of the Internet. As a business, we use most social media channels. The most successful for us are Google, Facebook, Instagram, and despite what I said about Twitter, it still works very well for us as a business.
What do you think is the biggest barrier to AI in commercial real estate?
Tushar: I think the biggest barrier to AI for commercial real estate is the body of data available. There’s a lot more data points in residential because it’s a consumer sector. There’s millions and millions of homeowners, buyers, and people renting space. Meanwhile, the volume of transactions in the commerical sector is much lower, therefore there’s less data points.
The biggest barrier to AI in general is how much data you can use and therefore
how many insights you can generate. If you were to take pre-existing data from the market and try to put it through an algorithm, it’s not enough to create any insights, which is why we’ve started off doing the work of getting listing brokers and automating their work using tech and data. Now, every time there’s an interaction between a customer and our platform, we pick up all of the data points. As a result, we can start building up our own proprietary data set and applying our algorithm and insights.
As with all data insights, the more data we have, the better our tools become. At the moment, we have relatively simple insights, but in five years’ time, our insights will have become very complete as the data compounding effect starts to take control. By then, more of the insights from our data will end up being more relevant compared to what you would ask the average office broker about the market.
Can you provide an example of a data-driven insight that a human being couldn’t have arrived at? And can you provide an example of something that the data arrived at which turned out to be wrong?
Tushar: At the moment, there’s a lot of examples where the data corresponds with what traditional property agents are saying, so for us, it means that we’re not far off. The fact that our conclusion align with that of traditional agents validates that we’re on the right track.
One example is what areas people are searching given a particular profile. For example, if you’re a PR agency and you put your email address into our platform, we generate about fifty lines of data. This data is collected and allows us to predict the location, office size, aesthetic, and price point you’re likely to be looking for based on the actions of similar companies that have used our platform previously. Our data shows that specific sectors are predictable, something which property agents can also conclude from experience.
Something that we’ve seen through our data, which is contradictory to popular belief is the amount of people that actually want coworking. There’s been huge hype around
coworking and how it’s going to take over everything and how everyone wants to be communal and share space with everyone else. Meanwhile, our search statistics show that over 70% of searches are for private offices. This tell us that despite the buzz around coworoking, people don’t actually want to be surrounded by other companies. They want to be in their own unit.
Sometimes, companies will end up taking fixed desks or coworking because the private
offices are too expensive and they can’t allocate their budgets effectively. That’s why the biggest thing we found, which we always advise coworking providers on is to allocate more desk space to private team space than they normally would have considered in advance.
As a portal, do you work on both the investment and the leasing side?
Tushar: One of the big opportunities for our data is to see what space is being used, where the demand is, where the supply and demand balance is, what size of units are being used, and what sort of customers are going into those units. If you look at a property information portal like CoStar or other commercial data portals, you’ll get good data around who’s a tenant of which building and how long their lease is.
However, as more and more of the market starts shifting into flexible space, building occupancy starts becoming more of a black box and it becomes difficult for landlords to understand the marketing. For example, they are only able to see coworking spaces such as WeWork or Second Home in a building rather than the business tenants themselves. It becomes difficult to understand what the rent rate is like, therefore which models they should adopt for their own buildings.
The information currently on the market is based on tried and true methods. That is, it shows that because some operators are successful, you should have similar facilities in your building to increase its valuation, whereas some operators are not, therefore you shouldn’t have similar facilities because it will decrease your valuation. What we’re trying to do build a dataset where people can value their buildings in a smarter way. For example, typically, anything with a serviced segment inside it has been sold at a discount because they don’t have a ten, fifteen year tenant, so we’re trying to bring more transparency to the sector.
Are you looking to sell solutions to larger companies?
Tushar: I think the majority of PropTech companies are trying to build software solutions to sell to large property corporates. We’re not one of them. We are we position ourselves as: if you were to start CBRE in 2015, Hubble is what it would look like. Rather than being a service-centric company as traditional property companies are, we’re built on a tech and data foundation and assisted by people. We’re essentially a new generation of property agents, so we’re not looking to sell our services to companies such as CBRE as we consider them our competitors.
However, the larger property companies we are looking to partner with are landlords, that is, people who own or manage office space. These are companies that are looking to understand what they’re going to do in the future, how they’re going to fill their buildings, and which tenants they should have. Our partnerships aim to help landlords fill their buildings while providing them with data insights.
You say you’re building “the new CBRE”. What kind of people work in these companies?
Tushar: We have a business of about twenty-five people. Half of us are software developers and a small portion of our staff are agents – in fact, we just hired our first one from Colliers about four weeks ago. He’s just in the audience there…he’s taken off the suit and is looking a bit scruffy these days so he might have to find a balance. The core composition of our company is quite different to who you’d see in traditional property where pretty much everyone is a surveyor and you might have some back office staff and office managers. We have a good split of backgrounds and most of our skills are digital skills.
How much real estate knowledge do people need to work in real estate? Is it purely technology and data?
Tushar: Obviously we do need a large degree of technology in the business. Myself and my cofounder are not from this sector, which helps us in the sense that we get to view it from a very different perspective – an almost child-like naiveite. This way, we get to see some problems as they are and are not jaded by anything. At the same time, there are a lot of bits that we don’t understand fundamentally so we make sure we hire in experts and have advisors around the business who can guide us in the right way.