Driving next-generation predictive business

Big Data Evangelist, IBM

Business prospers or perishes on a steady stream of smart predictions. And predictions can be life-or-death propositions for businesses. You want to know that demand for your products will be there or that you can source sufficient amounts of capital, labor and raw materials at acceptable prices before launching some ambitious plan. Moreover, you require trustworthy data to support your forecast that the product you’re developing will address a profitable, untapped niche when it comes to market. And you need predictive models, grounded in years of historical data across several business cycles, to corroborate your belief that factors such as costs and currency exchange rates will not eat up all your profits when bringing an offering to market.

Predictive model justification

The competitive penalty for making misguided predictions can be dire, which serves as a reality check on this approach. But even if the stakes aren’t quite that severe, you can’t inspire confidence in your business plans if you don’t justify them with high-quality predictive models. Consequently, tools, platforms and practices for building, refining and putting predictive analytics into operational business processes are the heart of any well-run, modern organization., predictive skills are beginning to permeate the business world at every level. Organizations of all sizes, in all sectors and geographies, are using these tools to drive evidence-based predictions into the full range of business processes, operations and decision points. And the technology has been eagerly embraced by a new breed of data scientists as their core power tools for sifting through big data to find historical trends, chart alternate predictive scenarios, drive real-world experiments, flag potential risks and identify new opportunities.

As I discussed in the blog post, “Decision confidence: Where the predictive chickens come home to roost,” technical best practices for predictive analytics span a wide range. However, they all come down to your ability to have business and technical professionals achieve ongoing consensus on the following critical decisions that are relevant to tuning your predictive models. These decisions all involve selecting the right: 

  • Predictive analytics problems
  • Subject populations for which predictions are to be made
  • Sources of data relevant to predictive problems
  • Data samples from which to build and train the predictive models
  • Data and model versions with the best predictive fit
  • Predictive variables
  • Predictive modeling approaches and algorithms
  • Predictive model-validation frequencies
  • Model-fitness approaches
  • Predictive visualizations

Analyst reasoning for predictive model adoption

If done effectively, predictive decision making can deliver quantifiable payback in organizations that put it into practice. Ventana Research, for example, is a good source of useful market research on predictive modeling as a critical business-technology competency area.

“Companies have been spending money on big data and visualization initiatives where software returns can be more difficult to quantify,” said Tony Cosentino, vice president and research director at Ventana Research. “Fortunately, the business case for predictive analytics can cite tangible business benefits, the most often of which our research identified are achieving competitive advantage (57 percent), creating new revenue opportunities (50 percent) and increasing profitability (46 percent).”

In a recent Ventana Research analyst perspective, Cosentino goes into greater depth on trends in adoption, staffing, deployment and management of predictive analytics as surfaced by their quantitative market research. Just as important, its findings clearly indicate that predictive analytics is a clear competitive differentiator in industries spanning today’s insight economy.

In a separate perspective from the same analyst in Information Management, Cosentino presents findings from Ventana Research on the extent of the skills gap that prevents many organizations from achieving the full benefits of predictive analytics. The research shows that success depends on mixing business and technical skills in cross-functional teams that are addressing predictive business challenges. Cosentino notes that it’s not common to find all the requisite skills in a single individual, which necessitates finding a diverse mix of specialists in business and technical areas who can collaborate productively on tough problems that demand creative multidisclipinary solutions.

And in yet another perspective from Cosentino in Information Management, he points out the vital importance of investing in the right tools to help drive team productivity on predictive analytics challenges. A key finding is that the success of any tool depends on clear consensus among technical and business stakeholders on key requirements, such as whether they are best deployed on premises or in the cloud.

Exploration into business transformation through predictive models

To gain further insights on how predictive analytics is driving business transformation, check out the executive summary for a recent Ventana Research study from which all these findings originate. And experience the power of this technology in your transformation strategy by starting your own predictive analytics journey today.