Blogs

Center your business strategy on predictive analytics at IBM Insight at World of Watson 2016

Post Comment
Big Data Evangelist, IBM

Successful businesses in the 21st century are those that drive predictive analytics both into their customer-facing initiatives—for example, through target marketing and next best offers—and into their back-end business processes. On 8 September, I participated in a CrowdChat in which subject matter experts and industry influencers discussed just this. Here are some of the high points of our conversation.

How can businesses make revenue-generating insights actionable?

Measure them to discover whether they actually contribute to revenue growth. Programmatize them using data-driven predictive models and decision automation. Evaluate, iterate, score, report—and repeat.

How can the combination of predictive analytics and big data help organizations gain a competitive advantage?

If you have a large pool of historical, profile, transaction and other customer data, you can build nuanced segmentation models that allow you to model and predict customers’ propensity to churn, accept offers and so on. In particular, frequent updates to customer data allow you to update your predictive model, helping you judge what customers are likely to do next—and that’s a revenue-boosting advantage when you’re driving next best offers.

Similarly, when you have a wide variety of customer data types (e.g., transactions, sentiment, clickstreams, social), you can obtain a well-rounded predictive portrait of customers’ propensities. Not surprisingly, then, a high volume of customer data—such as that gained from a deep historical record of customer transactions and interactions—can help you develop and validate predictive models of how customers will behave in a wide range of scenarios, allowing you to make robust predictions.

http://www.ibmbigdatahub.com/sites/default/files/predictiveanalytics_embed.jpgWhat skills and capabilities are needed to succeed with analytics?

Succeeding with analytics requires three things:

  • You must understand how you can use data, structured reports, visualizations, predictions and the like to enhance your decisions. What’s more, you must deliver insights from analytics to the right decision makers at the right times.
  • You must have a center of excellence that recruits, cultivates and manages the skills, tools and practices that drive your data mining, predictive modeling, visualization and much, much more.
  • You must infuse your organization with a passion for data-driven decision making at every level. Make it the foundation of your business culture.

How can the inability to forecast future events limit a company’s growth?

If you can’t forecast demand, then you risk not being able to deliver enough of the right kinds of products to realize target levels of revenue when demand materializes—a predictive weakness that can doom any business. Similarly, if you can’t predict important events—for example, those having to do with customer retention, offer acceptance and the like—then you might miss out on easy opportunities for revenue.

How your competitors might respond to various marketplace conditions, including your own moves, is of paramount importance in business. If you can’t predict this, then you risk handing the strategic advantage to your competitors. In much the same way, if you can’t predict marketplace disruptors and other “black swans”—or at least mitigate the risks of “unforeseeables”—then you leave yourself strategically vulnerable no matter how much your revenues are growing.

How can predictive analytics be expanded into different areas of the business to boost profitability?

Revenue growth is just one component of your overall financial results. Using predictive models, you can control costs and boost efficiencies by identifying trends in costs of labor, raw materials and the like. What’s more, your chief financial officer relies on predictive models when attempting to identify your exposure to likely fluctuations in currency exchange rates, borrowing rates, stock prices and more.

Yet that’s only the beginning. For example, only consider how your demand planning, materials management and inventory management professionals depend on predictive models to determine whether you’ll have enough raw, semi-finished and finished goods to meet likely levels of demand. Even your human resources managers look to predictive models when gauging rates of likely attrition in various geographies, positions and business units, allowing them to mitigate shortfalls through recruitment, incentives and the like.

Which predictive analytics techniques have proven particularly effective for growing revenue?

Essential techniques for effective prediction run the gamut of CRISP-DM. Indeed, what are you trying to predict? Do you have the right data? Do you have a valid statistical model? Have you iterated to ward off model decay? The list goes on. Accordingly, study the four pillars of predictive analytics—revenue growth in particular. Also, learn about effective steps you can take while doing predictive analytics.

Remember that the core applications for predictive customer analytics all focus on revenue growth: churn prediction, upsell prediction, cross-sell prediction, next-best-offer prediction and so forth. Not surprisingly, then, revenue growth is one of the outcomes you seek to predict as you search for practical things you can do in your business now to help ensure that your predictions come true. Indeed, sales, customer service and effective marketing all hinge on how well you can predict revenue growth. Similarly, demand generation requires valid predictive models of tactics—for example, events, content, ads and SEO—that are likely to prove effective.

Predictive business will be a primary theme at the upcoming IBM Insight at World of Watson 2016, scheduled for 24–27 October in Las Vegas. When you attend, be sure to check out the following sessions. To learn more, use the session preview tool to explore them ahead of time.

Advanced analytics

  • 2976: Predicting employee retention for data-driven human resources with IBM SPSS at Coca-Cola
  • 3634: Think you REALLY know your customers? With microsegmentation, you can.
  • 1675: Prescriptive analytics with decision optimization: An introduction for the line-of-business
  • 3587: What’s new in predictive and prescriptive analytics in the cloud

Don’t miss IBM Insight at World of Watson 2016, where you’ll be able to take part in the continuing evolution of the event once known as IBM Insight. Register today to attend IBM World of Watson 2016—we look forward to seeing you there!