When the small-market Oakland Athletics employed “moneyball” to make the American League Division Series in 2002, they changed the way baseball used statistics. Other teams quickly picked up the use of predictive analytics, and now it’s simply become the way things are done in baseball.
In the Brad Pitt-helmed movie about that team’s success, one character summed the power of the new model:
“Anybody who’s not building a team right and rebuilding it using your model, they’re dinosaurs. They’ll be sitting on their ass on the sofa in October, watching the Boston Red Sox win the World Series.”
Predictive analytics changed the way baseball teams choose their players, but there were nevertheless plenty of stubborn general managers who were slow to adopt the trend—and who were left watching other teams play in October.
The same is true in business, according to Jayne Landry, global vice president and general manager for Business Intelligence at SAP:
“Many executives continue to rely on gut feel over data and analytics. But we’re seeing that organizations using analytics extensively for business initiatives outperform their peers in terms of revenue generation, profits, and market valuation.”
It’s time to gain a better grasp of predictive analytics so your company isn’t the one left sitting on the sidelines.
Rules of the predictive analytics game
For the A’s to win all those games, they needed to know the rules. The same is true with predictive analytics: You need to know the basics. Simply put, it’s the use of historical data combined with present-day trends to predict the future. A little crystal ball-ish, I know, but predictive analytics uses data to find historical patterns and then uses those patterns to make informed predictions. When combined with your own knowledge about customers and your industry, it’s a tool you can use to make better decisions for your company.
Predictive analytics includes three main components:
- Data – Good data is everything. No algorithm or statistical analysis can make a good prediction based on bad or incomplete data.
- Statistical modeling – A variety of statistical models are used in predictive analytics, with regression being one of the most common.
- Assumptions – Predictive analytics assumes that the future will follow the same pattern as the past.
Computers do most of the statistical analysis these days, and it would be easy to confuse predictive analytics with machine learning (and sometimes those terms are used interchangeably). However, machine learning is just one branch of predictive analytics where computers are fed training data, and the machine “learns” how to interpret the data. Machine learning is useful in some forms of predictive analytics, but it is by no means the only form.
A home run for your business
Predictive analytics can help your business hit a home run. From predicting customer behavior to understanding when a machine will need maintenance to identifying what your employees need from their health plan, these valuable insights can change the way you do business.
Predictive analytics also allows you to personalize content to your customer, be more efficient with inventory management and identify high-value sales leads. The opportunities to improve your business with predictive analytics are almost endless.
Don’t strike out
Predictive analytics isn’t a magic wand that gives you all the answers, however. If you rely solely on it to make decisions, you’re headed for an out.
Predictive analytics is just a tool in your business toolbox. As I mentioned earlier, good data is essential for this tool to work. Things can get tricky as you continually run predictive analysis because for it to work well, you have to keep feeding those statistical models more and more data to keep the predictions coming. This can at times be frustrating and confusing as you try to figure out which data will give you the most accurate outcomes.
And like any tool, you have to wield the tool of predictive analytics wisely. Remember that any prediction you get about customer behavior, machine maintenance or sales leads needs to be measured against what you already know about your customers and your industry.
For example, a statistical model may not be aware of a recent change in the way your industry does business, or it may not have the latest data that your biggest customer was recently purchased by a competitor. Statistics and modeling can’t replace your biggest asset: your own knowledge and instincts.
Those 2002 Oakland Athletics rode predictive analytics all the way to the American League Division Series and changed the way baseball teams are created. Using predictive analytics can help you knock one out of the park for your company, too.
Mike Talbot is the vice president and chief technology officer at Veracity Consulting, a tech consulting team of problem-solvers and truth-tellers who deliver customized IT solutions for commercial and government clients across the U.S. Learn more at veracityit.com, and share your thoughts on Facebook or Twitter @engageveracity.