Ever since Facebook founder Mark Zuckerberg borrowed a mathematics concept called graph theory to coin the phrase “social graph,” it has become fashionable to classify Internet sites and mobile apps
by the kind of graph of data generated by their users.
Think of the “graph” as a grid or spreadsheet in three dimensions. Facebook’s social graph consists of a matrix that captures each individual’s relationship to friends and contacts, cross- referenced by their mutual interests, affiliations, associations, and locations. At the intersection of friends and interests lie the attributes that marketers crave most: passion, commitment, priority, aspiration, and identity.
The social graph provides another way to understand the value of the data generated by Union Square’s portfolio of venture-funded startups. The graph concept reveals the primary purpose of each company. For example,
> Foursquare generates a “location graph” that maps a user’s behavior by location.
> Kickstarter yields a “crowdfunding graph” of passion projects indexed by the people involved, such as creators, followers, and funders.
> SoundCloud generates a “listening graph” of musical compositions indexed across members who produce and listen to them.
> Twitter pioneered the “interest graph” that cross-references discussion topics by participating members and followers.
Each company in the portfolio presents itself as a unique marketplace matching those who seek a certain type of information with those who provide it.
The graph makes the market even more efficient by solving a problem that is unique to two-way digital networks. Every website and app that invites users to participate faces a fundamental problem of scale. That is, as the number of participants scales to millions and tens of millions, they generate so much content that it soon becomes impossible for the average person to navigate using traditional organizational principles like curated lists, indexes, and directories. Classic Internet search engines are often inadequate under these circumstances unless the user knows exactly what he or she is
seeking, and standard navigational concepts like categories and menus begin to break under the sheer volume of content. The solution is the graph. The graph is a new way to make the content navigable for all visitors.
The brilliance of the graph concept is that the data itself is used to make other data findable. As author David Weinberger remarks in a blog post called “Metadata and Understanding” on
KMWorld.com: “The solution to information overload is more information . . . so long as that more information is metadata.” Here’s how it works. The social graph cross-references one type of data with another to instantly generate a list that is relevant to a specific context. One topic becomes a lens that pulls the other information into focus. Naturally, depending upon the site or service being used, the lists vary by subject matter such as the most popular coffee shops sorted by location, recently published books with the most reviews, the blogs with the most followers, today’s most-shared articles, the crowdfunding projects with the most supporters, and so on.
These are not static lists or fixed navigation elements hard-coded by engineers: they are dynamic, which means that they respond instantly to the activity of other users on the site. The lists are not curated by human editors: they are generated automatically by algorithms. Continuously refreshed lists fed by real-time data have enabled Foursquare and Yelp to deliver superior results compared to a site like Citysearch. Whereas Citysearch relies on its editors to research, assemble, and publish
information about the businesses in each city—which means it is often out of date by the time it is posted—Foursquare and Yelp receive constant updates from their users.
There’s more. User engagement makes the service more valuable, not less valuable, than a site with professional editors. That seems like a paradox if we consider user-generated content as inferior to professionally produced information, but it becomes clear in practice. For instance, suppose you and I are both active members on a restaurant review site. We’re both foodies, and we love to
discover great new gastropubs and sushi joints. Your participation on the site makes the content more relevant to me, and vice versa.
The most active users automatically become navigational guides for their followers, whether or not they consciously choose to play such a role. Their activity on the site, such as posting comments, reviews, and rankings, will eventually pull a wide range of disparate content into focus, making it relevant for millions of others who join the site later. And conversely, even the most obscure content can serve as a way to discover like-minded users with shared interests, even if those other users happen to live on the opposite side of the planet.
The social graph could never exist on paper or in fixed media. It would be frozen and swiftly
outdated. Traditional manual editorial processes and library systems are incapable of organizing vast information flows in real time.
The generative approach pioneered by the companies that Union Square invested in around 2006 offers a sharp contrast to traditional businesses that produce information by hiring experts. In the traditional approach, the value of the content diminishes with time and use and is eventually depleted as more and more people consume it.
WHY TRADITIONAL COMPANIES MISS THE MARK
Movie studios, record labels, and book publishers invest huge sums in original content. So do advertisers and their agencies. It’s a risky business. Sometimes it works and the result is a big hit.
More frequently, this approach fails to deliver the expected economic return. Most traditional media businesses are vulnerable to the cycle of hits and misses, with a few big hits offsetting a string of misses. It’s impossible to make every film a box-office success or every book a bestseller. Even a hit factory like Pixar Animation Studios, which has produced such high-grossing, award-winning films as Toy Story, Finding Nemo, and Monsters, Inc., will occasionally release a dud.
In 2013 I was invited to address the executive team at Turner Broadcasting System. In the cable TV
industry, this team is the reigning heavyweight champion. Along with sibling HBO, the Turner channels deliver the majority of the revenue and nearly all the profit to parent company Time Warner. But these firms still operate on the old model of investing in content in the hope of making a hit show. My
mission was to shake up the team’s thinking a bit.
The topic of my session was “Thinking the Unthinkable.” Inspired by Brad Burnham’s criterion I asked my audience to compare Turner Broadcasting’s proprietary data asset, TV shows, with the data asset generated by social networks. Specifically, I drew their attention to the following comparison:
Pay TV Networks' Data Assets Social Networks’ Data Assets
Type of data Original TV production or licensed TV series
User-generated content such as photos, videos, posts, comments, status updates
Production
cost Expensive Free
Marketing and
promotion
Expensive Free (viral word of mouth)
Consumer fee
for access Expensive subscription Free Availability to
end user
Scarce (limited to certain
windows of time and territory) Abundant, forever, globally available Access by
third party
Exclusive access under distribution contract
Freely available to all via API and developer program
Life cycle of
data Quickly depleted Evergreen and continuously replenished
In my comments to the executives at Turner I used the television business as the example to make my point, but this comparison is hardly limited to media companies. It’s valid for any company that participates in the networked economy, including manufacturers and retailers and their advertising agencies. Every business relies upon media to communicate with customers, partners, suppliers, and employees, and that means every company needs to understand how the dynamics of media change when business migrates to the two-way network.
If your company produces information by hiring staff and investing in the production of data, the inevitable outcome is that you will perceive your data as expensive and therefore precious. You’ll strive to keep it scarce, locked in a vault. If, instead, you focus on investing in the tools and
infrastructure that enable others to generate data on your platform and make it findable, such as the application programming interface (API) that we considered in Chapter 4, you can reap an endless harvest of free data that can be shared, mingled, remixed, commented, replied to, and thereby compounded.
Every company that I consult with is aware of the opportunities in data but most of them confess to being overwhelmed. They find themselves swamped by the sheer volume of the data and they lack the technical skills to sort through it and make sense of it. They don’t have a history of partnering with others or sharing their proprietary data, which means they often can’t extract the full value of their data asset. As more and more parts of our physical economy are vaporized and replaced with digital data, massive new business opportunities exist in the collection, recording, organization, analysis, storage, indexing, and retrieval of information. Companies that streamline their business process around the collection and management of data will benefit by finding new ways to improve products, invent new ones, and keep their customers satisfied. These are crucial survival skills in the
Vaporized Era.
HOW THE MONEYBALL EFFECT CAN HELP
An increasing number of companies are finding ways to use the feedback loop of real-time data to optimize their business performance. I liken this process to the theme of Moneyball, the best-selling book by Michael Lewis that recounts the tale of a perennially losing baseball franchise that was saved by data. Aided by a young quantitative analyst, the head coach of the Oakland A’s began to track and study player performance data and apply the insights to optimize the lineup and improve overall team performance.
Moneyball isn’t really about baseball: the story is a hip allegory for management by data. The narrative includes a dramatic showdown between the coach and his old-school talent scouts who reject the data-driven approach, preferring to rely on instinct and intuition. Spoiler alert: in the end, the old-school dullards depart and the quantitative approach prevails, not just by transforming the team into winners but also by reshaping the management structure of the Oakland A’s. Ultimately the Moneyball approach has been adopted by the entire professional sports industry because once one team adopts a data-driven approach to optimization, its rivals have no choice but to respond in kind.
Rigorous quantification of professional athletes’ training, rest, nutrition, supplements, and performance has now become a universal norm.
Today companies in every industry have begun to adopt a Moneyball approach to data-driven decision-making:
> Wall Street bankers use automated high-frequency trading software to scan the market for buying patterns.
> Media buyers rely on automated bidding programs rather than gut instinct to buy advertising in specific target markets.
> Doctors in health maintenance organizations (HMOs) use checklists derived from broad surveys of best practices to treat patients.
All companies generate information. Whether you produce marketing data or how-to guides or retail SKUs, it’s a good idea to consider how you manage your data. Does your business create a proprietary data asset? Is it stored in a useable format or locked in a vault? Does that data asset grow or diminish with use? Does that asset grow or diminish with time? Do your partners contribute to its growth or subtract value? Does your business model depend upon the value of your information asset being depleted or constantly regenerated?
The answers to these questions will help you determine whether your company will sink or surf in the Vaporized Economy.