Real estate data is often heralded as a wildly valuable asset for stakeholders in the ecosystem. Data can inform the past, present, and future – it can be in pure dollars and cents form, factor into the acquisition or sale process, valuation, or financing considerations. It can be more operational in nature to understand tenant needs and wants. It creates measures of connectivity and sustainability, the condition of an asset over time…data is, quite frankly, anywhere and everywhere. But therein lies a problem – with so much data available, what’s useful to measure? How do you capture it? How do you capture it at scale? Does everyone define a metric in a similar way, so it’s always apples-to-apples? The short answers are that capturing data at scale, and more often than not, using data at scale, is a massive, yet lucrative, undertaking, that’s left many real estate incumbents and investors dumbfounded. With so much data available, why is it so hard? More tactically from a VC perspective, perhaps the question is “in a world where real estate data is so valuable, why is it so hard to create a massive data company?”. We believe the answers to those are complementary, but also somewhat challenging to explain without peeling back a few layers:
First, there is so much potential data in real estate that you should start with what you’re trying to solve for, then work back to what data you need. If you’re a landlord, your data needs evolve over the life of your involvement with an asset. You have data needs to inform your purchase price – location data, demographics, rent comps, all sorts of data to support a potential loan, data on the physical condition of the asset, maintenance records, utilities spend (for starters). Then, when you own, you need to stay on top of maintenance and tenants with assessments, roof conditions, data on turns and what is needed across your units (again, for starters), and with a potential exit you need to find buyers, and have similar financing and tax considerations. The data challenges for a single asset are high, and likely exponentially more complex at a portfolio level.
Assuming that you know exactly what data you need, we can move on to where you’re going to find it. Publicly available data sets can be purchased off the shelf, in a similar way that you could also pay for something like CoStar that will have a variety of data. But the more data you need, the more likely you are to ingest data from different sources, and in different forms. Comparing disparate data sets is a heavy lift (ask any analyst). Data also gets stale quickly – market comps from 2 years ago aren’t worth anything to you, but a comp from last week is immensely valuable. So now you need to integrate data that’s constantly changing, and make it useful. It’s getting tougher.
If you carry forward with this exercise, the picture that is painted is a messy one. Under a microscope, specific data can inform what to do about a specific problem, but when you zoom out, it’s a massive web of data entanglement and lack of structure. IT departments of landlords historically don’t get much budget, which is another contributor to the problem. Lack of resources to fight with the data problem doesn’t help.
A frequent dilemma that a landlord then faces is the traditional build-vs-buy question: if what I need doesn’t exist, shouldn’t I just spend the money now and try to build it for cheap? You could…but again, in a world where data gets stale and engineering talent will be needed on an ongoing basis, the biggest question we see is one of determining what core competencies of a landlord really are. In most cases, landlords are not going to attract or retain talent to build or maintain the software, nor will they be able to evolve with the changing needs of the business. COVID has highlighted this – as the demands of landlords has changed because of COVID, we’re seeing those landlords with more digital roots being much more nimble and responsive than those that have operated in an analog way.
In a ‘buy’ scenario, it gets challenging because the “Data & Analytics” market is so crowded. There appears to be fatigue in the market because landlords have been hearing about similar advantages from data companies for years, across different pricing structures and data availability. Sifting through the sea of options is time consuming and expensive. That said, some big companies have been successful in carving out a niche in the market – how they’ve done it is through a combination of grit and pounding phones (CoStar) to keep their data fresh, through gamifying the process (CompStak), and through first-mover advantages (more through establishing a trusted brand and comprehensiveness of data, as evidenced by the acquisition of 21 year old Real Capital Analytics by MSCI for $950M).
Venture-backed companies such as Cherre are also emerging – rather than acting as a source of data, they appear to be focused on making existing data actionable. There’s so much unstructured data in many organizations that Cherre needs to exist to integrate it, clean it, and make it accessible across the enterprise. It’s a timely solution to a long-standing problem.
The data problems in real estate have different solutions for different stakeholders: for a young startup targeting the space, it’s important to understand the status quo of the real estate data world. There are some go-tos that many people use (CoStar) but nobody likes, but it’s because they have comprehensive data that is always fresh. Without both, the durability of your business may be constrained to local market knowledge or for a specific period of time. Unfortunately, it’s tough to scale in a high-margin way unless you can create and sell ever-elusive proprietary data and not rely on people to generate it. For real estate incumbents, we encourage you to be an early adopter of data aggregation and cleaning platforms. Your firms will thank you in the long run, but it’ll be challenging to try to export data that is trapped in existing software and accounting systems. Without the effort, your data will continue to be locked away.
We wish there was a silver bullet to the real estate data problems. If you’re building a business model that breaks down data silos, structures data to make existing data more actionable, or automating the outputs that hours of data-ingestion and cleaning entails, we want to hear from you!