Data: Oil or Tulip Bulbs We are told that we live in the age of data. Improved collection and cheaper storage, combined with powerful new analytic technology like machine learning are transforming industries. Companies like Google, Amazon, Facebook, Uber and Tesla can apply powerful learning and optimization algorithms to their vast data sets, creating a competitive advantage that stifles competition and leads to monopoly. Every day, more data is collected in fewer hands through competitive action and acquisitions like the Facebook acquisition of Instagram and Whatsapp. Since data is the fuel of the digital economy, it is time for regulators to consider the data implications of mergers and consider breaking up data monopolies the way they once broke up Standard Oil. That’s one theory. The other is that the current mania for data is like the tulip bulb mania that nearly destroyed Holland’s economy in the early 17th Century. Speculators began paying outrageous prices for tulip bulbs and the resulting bubble made and destroyed fortunes. So what is the difference between oil and tulip bulbs? One is vital to the functioning of a society and the other is just an arbitrary store of value. Most people would consider data closer to oil because we have all had it beat into our heads that data is critical. But Dutch investors in 1637 considered tulip bulbs critical to that economy. Why does this matter? It pays to examine our prejudice towards considering data inherently valuable. First, data is not particularly rare. Most companies collect far more data than they need or will ever use. The vast majority of data goes unexamined and unanalyzed. It just clutters up servers. Further, most data is unstructured. Meaning that it doesn’t lend itself to analysis or application. Think of the world’s largest library, but the Dewey Decimal System doesn’t exist and all of the pages have been torn out of the books and thrown on the floor. But even if we have excellent, structured data and the tools and the will to analyze it, would that constitute a competitive advantage? Consider Google, a company that makes more effective use of data than any other company in the digital economy. Click through rates on a Google Display Network ad are around 0.35. Click through rates across the industry are about 0.3. Meaning that Google with their vast data stores and cutting edge analysis and optimization have beaten the industry by 0.05. (Yes, I know CTR is a crappy measure. But that’s still a quantitative difference without qualitative meaning.) Facebook with their lock on the social graph for most of the world is considered another potential data monopolist, yet Facebook has spent the last two years slowly admitting that ads on their network don’t noticeably outperform industry averages. The Googles, Amazons and Ubers of this world have good reasons for encouraging potential competitors to believe that their data constitutes an unconquerable competitive advantage. But I think most of their dominance is owed to good, old-fashioned network effects. If data is the secret sauce, I think we’d have seen stronger evidence by now. In a nutshell: Data is neither the tulip bulb or the oil of the digital economy. It’s more like dirt – common, cheap, and potentially useful in the right hands. Read More What should you do with data? While data in aggregate may not be especially valuable, your own company’s data often contains hidden value. The problem is that most companies collect data but do not have a data strategy that would allow them to weaponize that data for competitive advantage. Large companies have begun to try to address this missed opportunity by hiring Chief Data Officers (CDOs) who are tasked with determining how to collect, store, and analyze data across the enterprise. My own experience is that mid-market companies almost never apply the data they collect strategically. A recent article in Harvard Business Review entitled “What’s your data strategy?” is focused on the experience of large companies, but it’s lessons could be applied to companies of all sizes. Frequently, sales and marketing each have the responsibility to collect data that overlaps significantly in scope but does not agree in specifics. The HBR article points out that they lack a Single Source of Truth (SSOT). A company that cannot agree on how many widgets were sold last month and what source of traffic drove the sale of the most widgets is certainly wasting money and probably fostering interdepartmental conflicts. I have noticed that other companies have data strategies that define things too narrowly. Sometimes a customer is also a supplier, but if you only define them as only one or the other you have a limited understanding of your true exposure. What such a company lacks is a different kind of data strategy that features Multiple Versions of Truth (MVOT). In practice, an effective data strategy is a blend of SSOT and MVOT. Some data needs to be simple and clear, while other data needs to be parsed to reveal its complexities. Highly regulated companies tend to place more value on SSOT, while companies in competitive markets use MVOT. Why does this matter? The focus of this newsletter is on digital trends. But technology, particularly trendy technology can be distracting unless it is tied to fundamental business goals. Recently, I was talking to the COO of a mid-market company who was very excited because they were starting to use Salesforce. I asked him if he spent much time looking at Google Analytics. He regretfully said no, but was hopeful that Salesforce would help to correct that knowledge gap. I warned him that Salesforce, like GA, like Marketo, like SAP, will only give you back what you put into it. There is no software tool that will compensate for the lack of a data strategy. In a nutshell: Establish a data strategy before collecting data. Read More Revenge of The Everyone Filter bubbles are real. Too many of us use digital media to create self-reinforcing echo chambers of pleasing opinions and low-friction factoids. But no one’s filter bubble is perfect. And that is bad news for digital advertising. One of the enduring cliches of general market advertising is: “I know half of my advertising budget is wasted, but I don’t know which half.” Digital advertising, with its ability to target and test, was supposed to eliminate that “wasted half” by only showing your ads to the specific individuals in your target demographic and target market. Filter bubbles were the whole point. But recent research by companies like Adobe and media behemoth GroupM Global seem to indicate that broadening your audience beyond your core market actually increases the effectiveness of digital advertising and leads to greater sales. Why does this matter? Think of all the people you talked to today. Inevitably, some were older than you and some younger. Perhaps some of them were of a different race or religion or socio-economic class. Perhaps you even talked to friends or colleagues in a different geographical region. This is your practical peer group for today. This is your “everyone.” If your everyone is talking about Kendrick Lamar today, you are likely to check out his music, even if you are not naturally inclined to listen to music in that style. Advertising works the same way. Because few of us live in a community entirely composed of demographically identical human beings, we are influenced by a diverse peer group. Advertising creates echoes through culture but those echoes quickly die if our practical peer group hasn’t seen the same ads. In a nutshell: Forge a balance between target market and general market. Read More