If you have ever bought or sold a home, you are familiar with real estate comps. All real estate sales have to be reported to the government so your broker can simply enter an address into any one of a number of databases (e.g., Zillow, ACRIS) to get a list of similar nearby properties and what they sold for.
Leases are different. There is no reporting requirement for leases and there are no comprehensive databases of all lease transactions. Your broker has to rely on whatever data his company has collected on the leases his company handled or ask other brokers to share their comps with him. As you can imagine, the comps he gets are incomplete and time-consuming to collect. CompStak hopes to fix all that.
CompStak is a give-get system. Brokers enter the lease comps they have in exchange for points they can use to get the comps they will later need. They can also buy points if they need more comps. CompStak will also sell this information to non-brokers for whom comprehensive, real-time data could be extremely valuable (e.g., hedge funds investing in real estate).
CompStak sailed right through my first cut. It targets a large but neglected market that is in many ways still stuck in pre-internet practices. And at first glance they are trying to replicate something that already transformed a similar market (viz., real estate sales).
I’ve never worked in real estate so my first step was to network to and talk with a lot of commercial real estate lease brokers. Their reaction to CompStak’s value proposition? They want it but doubt it at the same time. Here’s why:
- Data collection and integrity
- Incremental value of the data set
The brokers I consulted were unsure if many brokers would share their comps this way. Landlords generally require confidentiality clauses so sharing comps in a way that could be traced back to the broker could result in lost business or even litigation. Also, many large brokers consider their internal database of comps to be a competitive advantage and have a blanket policy of not sharing comps. Furthermore, it is of a pain to enter comps; some large firms insist that their brokers enter the comp in their system before paying them for that transaction to insure compliance. To avoid this extra effort, many brokers might simply buy points as needed or pay a flat monthly fee for unlimited comps (if given the option) rather than enter their comps in to CompStak. And as if that were not enough, the data that does make it into the system could be suspect; landlords might try to game the system by entering inflated lease rates or understated concessions (e.g., buildout allowances) so as to influence future leases.
CompStak’s founder, Michael Mandel (@brokerednyc), says that many comps are added by 3rd parties such as brokers that the landlord’s or tenant’s brokers informally shared a comp with, entities that are considering purchasing a property and who received data on all its leases as part of the due diligence process, etc., which might get around confidentiality clauses. At the same time, CompStak tries to get multiple comps for a particular lease to ensure completeness and accuracy. It’s also worth noting that CompStak is currently in beta and Michael believes he already has comps for over half the lease transactions in Manhattan. His success so far partly addresses my concerns on data collection. That said, only time will tell if they continue collect enough accurate comps after the initial “let’s give it a try” phase passes.
The second set of concerns revolves around the value of the data. For brokers, comps are only the start of the process. They are useful for setting clients’ expectations but at the end of the day, what matters most is what competing offers are actually on the table. Comps may have minimal impact at this point. Granted, you could make the same argument for residential sales but the brokers I spoke to felt that the greater complexity of commercial leases (there are many more factors than just price per square foot) meant that the lease offers they receive are less informed by comps than as is the case in residential sales.
For non-brokers, the question is how much incremental value CompStak can provide over existing data sources . Most properties for lease list their availability on CoStar. Since landlords remove properties as they are filled, at the very lease CoStar has some data on the lease and the last asking price. Combining those quasi-comps with historical data on discounts off asking in the final contracts for which they have data (note: CoStar has recently been trying to collect final contract data to turn these into true comps) and adjusting for recent trends in asking prices in that market could (in theory) get a real estate investor pretty far. At the same time, firms like CBRE Econometric Advisors (formerly Torto Wheaton) provide detailed market outlooks and forecasts for the most of the large markets. Since they are owned by the largest commercial real estate services firm and presumably have access to their internal database of comps, their data is based on a significant representative sample of the lease market. While their more detailed reports come out quarterly, they also offer “strategy services” for substantial clients who need more detailed or real-time data. The big unknown is how much additional accuracy and/or timeliness CompStak can provide. If real estate investors do not come to perceive CompStak’s data as substantially better than what’s available, CompStak will be seen as just another data point and may end up competing with established data sources on price… which is generally not a happy place to be. My sense is that this second issue will ultimately depend on how well CompStak does at meeting the first challenge.
So going into my second cut, it became clear to me that CompStak will come down to market acceptance. If CompStak rapidly amasses a large enough share of the transactions in a given market and these comps are reasonably accurate and complete, we’ll see if the incremental value of more comps and time saved collecting comps is enough to get brokers to change their current behavior to make entering their comps into CompStak part of their everyday routine. As is often the case, even the end users are not sure if this will happen.
For investments like this, you have to go with your gut and without any first-hand experience in real estate, I simply don’t have a gut feel for CompStak. If there were someone I knew with a solid track record in startup investing and relevant domain experience to back up his gut feel who was gung-ho on CompStak, I might go along for the ride.
(Just to be clear, I am not a fan of outsourcing my investment decisions but in certain very specific circumstances, where the judgment of other relevant investors informs gaps in my analysis and can be the factor that tips the balance.)
Feel free to agree, disagree, or insult my investing skills Limerick style.
Update: you can view SpokenLayer’s pitch at TechCrunch Disrupt here
What can I say about SpokenLayer? I want this service yesterday. My only question is if this is a viable business.
Full disclosure: In 2005 I spent several months thinking through a business remarkably similar to SpokenLayer.
Let’s say you are driving to work, running, in the shower, or in any other situation where you want to read your favorite newspaper but either cannot or prefer not to read. SpokenLayer will read it to you.
The concept itself is seemingly simple but actually pretty powerful. With SpokenLayer, any provider of written content can now instantly become its own radio station. The publisher gets a new channel to its audience and a completely new vehicle for advertising, with virtually no additional investment of money or other resources.
SpokenLayer automatically converts an article’s text to speech. Quality is pretty much what you’d expect for an automated text to speech program but for short-form, perishable content, this may be good enough. Additionally, they will manually narrate specific news articles as they become popular or as requested by their clients. They also give the authors of the articles (and possibly even readers themselves) the option of submitting a narration.
In addition to revenue from manual narrations, the main revenue model is audio advertising, split between SpokenLayer and the publisher.
The problem is what if they succeed? What’s to stop their largest publishers from in-sourcing this service? Creating a multi-publisher platform, with publisher-level and topic-level pronunciation guides, and an app with great UX is certainly not trivial. But hacking together an ok-but-not-great text to speech engine to glue onto a single publisher’s app is at most a few months work. If SpokenLayer takes off and they start generating significant ad revenue, their largest customers will have a very simple make vs. buy calculation. SpokenLayer will have to slash their share of ad revenues to the bare minimum to make it not worth their while to in-source the service and even that may not work.
- Aggregation: The aim is for SpokenLayer to become the one-stop-shop for this kind of audio content and that as it captures more and more content, network effects will kick in. Publishers who want their content to be discovered will have to be in their network. To me, this feels like a sonic twist on the old portal model… which we know is just kicking ass for Yahoo, AOL, et. al. Kidding aside, the content discovery angle may be valuable for the smaller publishers but the larger ones are already one of the few always-go-to news sources for their audience. So if the New York Times decided to pull out of SpokenLayer and embed text to speech in their own app, their readers would follow them… with is exactly what happened when they cancelled their deal with AvantGo and created their own mobile site and app.
- Curation: There is no doubt that there is a lot of mediocre and irrelevant content out there so there would seem to be some value in creating a Pandora-like stream of relevant articles to listen to. The question to ask is this: Aren’t the publishers already curating this content for us? I read the New York Times World section at least in part because they do a decent job of selecting news that interests me. For SpokenLayer to add value, it would have to do a good job of finding articles that the Times missed or different takes on the same topic; serving up similar articles on the same topic would simply waste my time.
- Specialization: Simply put, it’s difficult and expensive to develop new technologies. On the other hand, the large publishers have the resources and have shown that they are willing to deploy them if there’s enough money on the table. Plus, the lag time between new & hard to commonplace & easy is getting shorter every day….
- Owning the ad network: Creating the audio ad market is not just hard, but requires scale. Once you have it, you have enough impressions to offer up very specialized content and command higher CPMs. If SpokenLayer gets to scale first, this will give them a solid cost advantage against competitors. But again, the largest publishers already have this scale and, if they don’t already have a radio division, can (in theory) cross-sell their current advertiser base into audio ads.
Even though I am not entirely sold on Will’s argument, I put quite a bit of time into understanding the opportunity. If Will makes a strong case that even with just the middle and low end of the market, there is enough there to make SpokenLayer into a big company, I would likely put it even more time getting to know his team, talking to his partners, advisors, and other investors, digging into the terms of the deal, et. al. But since the point of this blog is who passed my first cut (and why), the bottom line is that SpokenLayer made it at least that far.
Whether I invest or not, I want SpokenLayer to succeed… even though it would make me feel pretty dumb for passing on the idea in 2005. 🙂
Feel free to agree, disagree, or insult my investing skills, Yo Mamma style.
AngelList profile: http://angel.co/spokenlayer-1