Predictive analytics for the small business – part 1

1.smallThe Big Data revolution continues to transform all levels of business and society in ways almost unimaginable just a few short years ago. Indeed, the best word that comes to mind in describing this phenomenon is “epic”. The McKinsey Global Institute in 2011 published a 143-page report on Big Data, proclaiming it as “the next frontier for innovation, competition and productivity.” The report notes four major eras of IT adoption over the past 50 years: mainframes (1959-1973), minicomputers and PCs (1973-1995), Internet and Web 1.0 (1995-2000), and Mobile devices and Web 2.0 (2000-2006).

 

 

image

 

 

According to McKinsey, Big Data will amount to a fifth wave in the technology revolution and lead to major surges in global productivity. In the U.S. retail sector alone, for example, it is estimated that big data could increase a retailer’s operating margin more than 60 percent.

 

Closely aligned and integrally related to Big Data is the field of Predictive Analytics, or PA. As one article declares, predictive analytics is “Big Data’s Greatest Power.” The article goes on to describe PA as the science of taking the vast store of historical data – the ginormous amount of structured and unstructured data stored “out there” – and using data mining and statistics to make accurate models and predictions of future customer behavior and business scenarios in near real-time.

 

Personal data is now considered a new economic “asset class” much like gold, silver, oil, and precious metals. And for businesses of all sizes this obviously implies a wealth of opportunities. Just like websites and social media in the last decade, and mobility and cloud technologies in recent years, getting onboard with Big Data and predictive analytics is critical for business success in today’s ultra-competitive landscape.

 

 

image

 

 

Data is ubiquitous and is being mined from increasing numbers of sources such as emails, payments, log data, telephone conversations, audio, video, social media, call center transactions, sensors, RFIDs, and many, many more. Considering that the amount of data today is just the tip of the iceberg of what will be available in the years to come, the sense of urgency around building business application that can mine this data becomes even more paramount.

 

Big Data and predictive analytics are integrally related and crucially important to business and commerce today. The exponential growth of Big Data and the power of predictive analytics are combining in altogether new ways to provide unique and almost unheard of opportunities for business growth. PA can be used for everything from anticipating which customers are likely to defect from a company or cancel a subscription, to what kinds of merchandise people will purchase and products they will choose, to insurance fraud detection.

 

Everyone gets junk mail. Well, imagine as a business being able to predict which of your customers will most likely respond to those mailers. That’s what PREMIER Bankcard did. By using targeted marketing, the bankcard reduced mailing costs by $12 million.

 

Now the “big question” when it comes to Big Data and predictive analytics is how exactly to get started, especially if you’re a small business without lots of resources. This article outlines “nine steps to predictive success” and provides a great starting point for business and IT shops to start thinking about how to implement a PA strategy.

 

There are lots of different recommendations and strategies out there about how to implement predictive analytics, but here are the core points you’ll want to follow:

 

  • Business requirements: Understand the business problem you’re trying to solve.

 

  • Data understanding: Ensure you have enough data and know the history behind it.

 

  • Data preparation: Raw data is collected, cleansed, and prepared for predictive modeling.

 

  • Modeling & model evaluation: Modeling techniques are reviewed; model is built and evaluated to see if it meets business goals.

 

  • Deployment: Apply your predictive model; consider a data visualization tool to showcase it.

 

Please join us in Part 2 of this series as we begin to explore these guidelines more fully. In the process, we’ll learn what it takes to kickstart your predictive analytics strategy in a way that will set your small business apart from the rest of the competition.

 

You might also like