Some of you may have listened to President Obama’s State of the Union address last week. You might have noticed his emphasis on manufacturing and wondered why. David Brooks did. In his New York Times op-ed following the speech, he note that “Ninety percent of American workers work in the service economy, but Obama spoke mostly about manufacturing.”
The U.S. economy is a service economy.
By extrapolation, if you’re a CFO reading this, it’s highly likely you’re the CFO of a service business. But what is a service business? Is service answering the phone nicely? Is service flipping burgers? No. Service is delivering information to you that is personal and relevant.
Say you’re staying at a hotel. You ask the concierge for a cheap Chinese restaurant within walking distance. If you’re steered to a good place, if you get the right information, that’s service. If you go to a doctor and she says that based on your personal genome and your lifestyle, you need to start taking Lipitor and, by the way, it wouldn’t hurt to start exercising, that’s service.
Service in the Digital Economy
Now look at the Amazon.com website. What Amazon is trying to do is deliver information that is personal and relevant. People like you bought this book; they liked it; maybe you’ll like it, too. Now look at the website and locate the transaction processing system. It’s that little shopping cart in the upper-right hand corner. It’s pretty hard to find. What does that tell you about how important it is? Yes, it has to be secure; it has to be available, and it has to work, but it’s really not that important because Amazon is about service, not sales.
Up to now, most of the work in the IT industry has focused on transaction processing, whether that’s automated teller machine (ATM) systems or implementing an order-to-cash business process flow. Yes, they have to work, but how important are they?
So what does this mean if you’re a healthcare provider or a financial services company? In short, what does this mean if you’re running a service business? You may never have thought about this, but the surface Web, the Internet that Google lets us search, about 4.2 billion pages averaging about 320 kilobytes per page, is around 1,344 terabytes. By contrast in 2010 disk drive manufacturers shipped over 5,000,0000 terabytes of storage – in one year. So, chances are, your company is already storing way more than Google’s terabyte load, but are you even scratching the surface of the potential of that data to deliver service? Unlikely.
Go to your favorite banking website. The only reason you login is for security purposes, so the system can grant access, not so it can deliver information that’s personal and relevant. From the moment you login, you’re interacting with a transaction processing system. You can debit, credit, purchase or sell a stock, but could the system tell you that people like you bought a particular stock today, or people like you refinanced their mortgage given falling interest rates? And that’s just the simple stuff, the stuff that Amazon has been doing for quite some time.
I’ve never liked going to large box home improvement stores, but consider this scenario. Four weeks ago, you bought some white Cararra tiles; three weeks ago, you bought a vanity; last week you bought a sink. Can you guess what’s going on? Well, of course: you’re remodeling your bathroom. And if over the past six months you’ve spent $25,000 at the store, why isn’t the store manager sending you an email or an IM telling you that a certain toilet is available today in Carrara White at a 10% discount and you can pick it up with her compliments.
Again, that’s service, and it’s easy.
The Barriers Are in Your Mind
The challenge today is not a lack of information. Your company, every company, has at least ten Internet’s full of data. The challenge now is to make that information personal and relevant.
The first thing you need to realize is that you’re a victim of the SQL hammer. If having a hammer means everything is a nail, then it’s the same for SQL.
SQL is a programming language, developed in the 1970s, used to populate and query large databases; that is, store and retrieve data. It’s a terrific tool for manipulating precise information, which makes it good for transaction processing.
Just to help you see how problematic this is for today’s challenges, imagine that Google’s search engine (which is trying to deliver information that is personal and relevant to you) was built by a bunch of SQL engineers. First, they would have designed a global data schema for all the information on the planet. Completing that task, they would have used extract, transform, load (ETL) and data cleansing tools to bring all the information on the planet into their global SQL database. Finally, they would write reports like, “Places to camp in France,” or “Chinese restaurants in Hickory, NC.” After ten years and tens of millions of dollars, the team would probably have given up. Fortunately, Google didn’t take that approach.
The New Big Data Hammers
So you need new hammers. The whole conversation about the consumerization of IT is taking you in the wrong direction. Sure, debating whether your employees should or shouldn’t be posting to Facebook at work, or how to block access to Dropbox, are warranted, but the bigger discussion should be about how you bring the technologies born in the consumer internet and apply them to the challenges of delivering information that is personal and relevant to your customers and suppliers.
Ask your CTO to give you a lecture on Hadoop, Lucene, Hive, Mahout, and Cassandra. These are technologies being used by the consumer applications to make sense of and deliver super large volumes information (Big Data) that is personal and relevant to you. Start a simple project that uses one of these hammers in order to build a service system for your customers and suppliers. Doing so will make your company’s State of the Union truly strong.
Timothy Chou teaches cloud computing at Stanford University. He is the former president of Oracle On Demand and the author of Cloud: Seven Clear Business Models.