by Kate PavaoMarch 2014
Big data is more than just big hype, says Thomas H. Davenport, author of Competing on Analytics, Keeping Up With the Quants and Big Data at Work, published in February 2014.
“People tend to not like the name very much,” he tells Profit. “But ‘big data’ calls people’s attention to some forms of data that they might not have been using in the past — external data, internet data, genome data, etc. The real issue, however, is how you combine that with the other types of data that you already have to make more sense of what your customers want, what your employees are up to, and what the demand might be for your products and services.”
To create a successful big data strategy, Davenport says executives should focus in one of four areas: saving money, making faster decisions, making better decisions, or creating new products and services. Here, learn more about his advice for launching a successful big data initiative in 2014 – including in which department you might want to start.
Profit: You consult with a lot of different companies. What kind of culture do you see at those that use big data successfully?
Don’t just wait for the data to be right before you start actually applying it to your business problems. It will never be
Davenport: You need an experimental attitude and the willingness to tolerate ambiguity. It’s kind of the same attitude that you’d have to have as an executive when you were in, say, a pharmaceutical company, and you have these drug development projects underway.
You may fail to create a great new data product the first time you try, or maybe you don’t actually improve the decision, or maybe you find out in cases that the great new marketing idea you had doesn’t actually seem to be working. But that’s pretty useful knowledge. It’s important to admit your failures and learn from your failures, and not worry about it too much, because it’s an inevitable aspect of experimenting and discovering with data.
Profit: On the flip side, what are some of the big mistakes that you see company leaders continuing to make when it comes to big data?
Davenport: They spend too much time and effort on getting the data perfect. A lot of organizations feel like they have to invest a huge amount of time and energy filling up their data warehouses with all the right information, and making sure the data quality is really great, and making sure that all the relevant data is integrated and so on. The problem with that is you don’t really deliver any business value at that stage. It’s all laying the foundation.
At most organizations, people are going to get rather impatient. I’ve seen a number of examples of executives saying, “Forget it, we can’t afford this anymore, we’re going to end this project.”
It’s very important to deliver value along the way, improve a decision, or introduce a new product or service. Don’t just wait for the data to be right before you start actually applying it to your business problems. It will never be perfect.
Profit: If executives want to pull off a big data project in 2014, where do you suggest they start?
Davenport: Most people have been very focused on internal data, but often there are opportunities to say, “How can we do a better job of improving our decisions or our products by augmenting all that internal data with some external data?” Can you scrape something from the Internet? Can you look at social media? Can you put a sensor in your product to collect some data about it? Or, there are, of course, tons of data brokers who have external data about your customers and more.
Profit: If you’re looking for some easy wins, are there particular lines of business that you would suggest executives focus on?
Davenport: Marketing is the obvious candidate because there’s so much data these days. It’s a very exciting period in marketing, but a very demanding one for marketing executives to change their orientations and become much more oriented to data and analytics.
At the macro level, you can say, “Are we spending money in the right places? Are our promotions working or not? Should we be advertising on television?” You can now actually address all of those things with big data and analytics.
At the micro level, you can start to say, “Is this version of the website better than that version? What is most responsible for somebody buying that product? Is it the last visit to the website, or the billboard they saw on their way to work?” It’s hard to do that kind of attribution analysis, but it can be done.
Photography by Shutterstock