We all look for a proprietary competitive advantage when we invest. If a startup doesn’t have a “secret sauce,” why bother? Since the collapse of the Internet bubble in 2000, the flavor of this sauce has been overwhelmingly technology-that is, companies with a proprietary technology that gives them a competitive advantage. Another time-proven flavor is the category of companies whose competitive advantage comes from a proprietary body of information.
The more unique and proprietary that information-and the more accessible, searchable and malleable its form-the more valuable it becomes for investors, entrepreneurs and customers alike. Like fine wine, information collected and stored within proprietary databases ripens with age.
Yet, at their earliest stages, proprietary database companies often aren’t seen this way. They are difficult to build, well disguised in their value propositions and sometimes subtle in terms of their competitive advantages.
Granted, at the seed-stage level, one often has to squint hard to see true business opportunities in front of what can often be multi-year development endeavors. Evaluating proprietary information companies requires that unique blend of due diligence, market opportunity, critical mass and often one large leap of faith in order to choose which of these mass repositories of data and information will actually succeed. Google is the obvious standout as a searchable proprietary database to end all databases. ChoicePoint is another high-profile example, though one where the value of all the personal private information collected for background checks may have become a bit too valuable.
Still, there are startups we see every day that hold the promise of combining highly valuable information within true proprietary databases yielding opportunities that are often difficult to foresee. By combining highly valuable information with highly proprietary databases, where the information is generally difficult to come by, great businesses can emerge and evolve.
We’ve seen everything from BeatTheTraffic.com (a unique startup that’s collecting, collating and organizing traffic data from 12 different markets and 4,000 data sources into usable, searchable real-time traffic information) to ETWaterSystems (which has a database with millions of pieces of information about plant species, landscape-specific watering requirements, irrigation systems and water district regulations). Each company uses its own applications and algorithms to marry its info with real-time data to save customers time, money, energy and even precious water.
Keys To Success
So what are the keys to success for these types of investments? In my 16 years of investing, some elements of a framework emerge. First, there is the concept of the marketability of the information itself. Marketability involves critical mass (sufficient breadth and depth), accuracy and timeliness. Think of a dictionary. One with words starting only with the letters A through M is not very useful. Similarly one that only lists words without definitions is equally lacking. And if only 90% of the spellings and definitions are correct, we’d probably have the equivalent of a “market failure.”
Yet, since dictionary information is static and not volatile, timeliness is not a critical variable. Take the example of BeatTheTraffic.com. If its information is 24 hours old, it is unusable to customers. In assessing the value that a startup proposes to deliver, the proprietary nature of the information, its timeliness and the lack of any other “good enough” substitutes therefore become the primary gating factors.
Suppose we have an entrepreneur with a proprietary information database that looks interesting. Is there also a business model that fits, can it make money and, further, could it yield venture capital type returns for early investors? On the expense side, as early stage investors, we must ask ourselves what it takes to build such a store of proprietary information and the ongoing cost to maintain it. The revenue side is the trickier challenge. Is there a market? What are customers willing to pay? How much of that information will they buy? How many potential customers are there? And how much will it cost to get them to show up?
We’ve watched with growing interest as Snap, an IdeaLab company, logs tens of millions of pieces of click stream data every day to improve search results by determining true Web user behavior. Yet, it is Snap’s algorithm applied to the proprietary data that powers its search and differentiates it.
Likewise, we have helped fund Savage Beast Technologies, a music “genome project” that allows music buffs to be matched with songs that correlate to their very own “DNA” of musical tastes. Savage Beast uses nearest neighbor matching algorithms applied to its proprietary database of analyzed music to issue custom recommendations. It commercializes the process by doing so on a mass consumer basis.
Hard To Copy
The common theme here is that the information must not only be difficult to obtain but even more difficult for anyone to easily replicate the collection, application or handling thereof. It then must be set within a platform where its use and value can be mined in myriad ways and launched in markets where customer demand is obvious and growing.
The choice for VCs becomes not just one of technological capability, management team and market size, but of the possibility that, with enough leg work, time and capital, we can squint just hard enough to see business opportunities emerge while maintaining the inherent competitive advantages such databases provide. Whether it’s traffic search or music discovery, the organization and management of all that information must add up to such an advantage that no competitor could economically attempt the same thing, and no customer would dream of living without that info.
When evaluating proprietary database deals, time is often the best due diligence tool. When we first saw Savage Beast nearly two years before we invested in it, its plan sounded far too challenging. The founders saw an opportunity to catalog music not by artist or album but by “genomic” traits peculiar to individual songs. It was clearly a difficult and expensive proposition to break down each of those traits, pull them out one by one from thousands of songs, and catalog them into large searchable databases. Yet, once that information was cataloged it could indeed prove highly valuable, a next- generation version of collaborative filtering at the song level.
“We were interested in having trained musicologists analyze an enormous number of songs based on anywhere from 150 to 350 “genomes” per song,” says Joe Kennedy, CEO of Savage Beast. “It was a way of understanding and matching music to consumers’ tastes, as opposed to a simplistic recommendation system some of the consumer sites were offering.”
No one had tried it. And we certainly weren’t convinced it would work. Yet, two years later Savage Beast was well on its way to cataloging every song on the Billboard 500 since 1957. It had deals signed with major customers, including AOL and BestBuy, and proved that even as costly as it was to build and maintain such a database that it could achieve critical mass. Having seen the progress achieved even in the absence of any institutional funding, we took that extra leap of faith that many proprietary database companies often require. In this case, time worked to our advantage.
Yet, early stage VCs should be careful that time does not also work against them. In a separate case, we had evaluated another company, Corvigo, an anti-spam company in the crowded security market that was also pursuing its own niche within the world of proprietary databases. Though it was a strong company with products easy adoptable for customer testing, we wondered whether the product was truly differentiated enough to fend off competition. As we continued to weigh our analysis, Corvigo was not only funded by Sequoia, but was acquired by Tumbleweed.
The lesson: Though time can often be the best due diligence tool, waiting too long often allows someone else to squint that much harder and find the true value hidden within the Googles, Snaps or Savage Beasts of the world.
If it’s a question of critical mass on the front end, it’s also a question of market timing on the back end. With Savage Beast, the digital music thing was supposed to happen five years ago. Fortunately, it took about that long for the iPod craze to catch fire just as the company was hitting its stride. Granularity of music distribution has shifted from the CD to the song. Combined with the rise in awareness and popularity of independent artists, the confluence of these trends raised the ante on music search and discovery modalities.
Snap now faces this same challenge of serendipity. “With Snap, it’s about getting success in search vs. being sent down a dark alley,” says Tom McGovern, Snap’s CEO. McGovern claims his company can “disambiguate” search results by determining user behavior. How? By collecting that critical mass of click stream search data after having examined millions upon millions of pieces of Internet user click-through behavior, as opposed to the word frequency data used in Google’s algorithm. Snap’s click stream data are processed through a file system Snap created, such that the data can be collected, organized and handled in real time. The critical mass of information increases in value over time.
It’s the essence of achieving critical mass that makes these companies not only workable, but also fundable. And yet there exists a funding paradox. Critical mass is great. Customer validation and traction of the application that rides upon the data is great. But these proof points attract interest of other, sometimes larger and later-stage investors. Perhaps even more than in purely technology driven deals, the confidence and quickness in pulling the trigger at the right inflection point is critical. It may be the most critical part of investing in early stage proprietary information-based companies.
Larry Kubal is a founding partner of Labrador Ventures, a Silicon Valley seed-stage venture fund. He represents the firm at portfolio companies Aperto Networks, Proficient Systems, Savage Beast, Soligence and Teraburst Networks. He may be reached at