Predictive forecasting for the cognitive era


When people think of the word forecast, the first thing that comes to mind is usually the weather. But in business, forecasts are produced for many reasons: product demand, sales, revenue, hiring needs and much more. A forecast is really a prediction about what will happen at some future date. If you look up forecast and predict in the dictionary, you’ll find that the two words mean essentially the same thing.

The ability to accurately predict what is likely to happen at a point in the future, and build plans and strategies based on that knowledge, is essential to an organization’s success. But what happens when a forecast is inaccurate? What is the impact on a business, its customers or its partners? For businesses, the ability to catch even a tiny glimpse of what the future may hold can lead to happy customers, improved efficiency and productivity, and highly successful business decisions. As a result, predictive analytics is changing the game for finance organizations.

Honing the accuracy of forecasts

On-the-go business professionals need accurate forecasting in today’s hypercompetitive business environment. Ongoing reforecasting helps managers to boost predictive accuracy that can answer these critical questions: What did we expect? How are we doing against our plan? And, even more important, how should we adapt our plans going forward?

For example, if revenue forecasts are below targets, a bank or financial services company may need to recalibrate products or services to attract new customers or keep current customers from leaving. With a model-based approach to forecasting, marketing can perform what-if analyses to test new product or service initiatives that examine impact by customer and customer segment. In turn, bank sales team members can evaluate these scenarios to adjust their sales strategy, such as maximizing time spent with the most profitable customers.

Updates to plans feed directly to the finance team. That team then turns the marketing and sales projections into net revenue projections—all in a matter of hours or days rather than in weeks or months, when remedial action may be too late. However, with data volumes growing exponentially and the business environment growing in complexity, traditional forecasting techniques born of spreadsheets may fall short in giving you the precision and insight you need.

Leveraging forecasting tools

To help boost forecasting accuracy in the era of cognitive computing, business analysts require tools that can go beyond the analysis of structured numerical data. Forecasting tools need to leverage advanced cognitive algorithms, predictive modeling and statistical analysis. They need to include information from unstructured text, including internal emails, customer feedback and even comments from social media. And tools need to include external factors in plans and forecasts, such as seasonal fluctuations or variations in sales volume attributable to customer demographics. Tools also need to surface new insights through mobile capabilities.

IBM Vision 2016, 9–12 May 2016, hosts the session, FPM-1261 – Predictive Forecasting for the Cognitive Era with IBM SPSS. In that session, Steve Barbee, offering manager, IBM Predictive Analytics Algorithms, at IBM, and I explore the latest features of IBM SPSS Modeler. SPSS Modeler, in conjunction with IBM Cognos TM1, brings big data analytics to the desktop and to cloud computing. Hear how big data algorithms can take forecasting to the next level by allowing you to model the ways that forecasts affect one another. See how to forecast with time and place in mind, and take advantage of the latest in advanced analytics to make smarter decisions tuned to a better future.

In addition, don’t miss the lab session, Hands On with Predictive Forecasting: Creating More Accurate Forecasts with IBM SPSS Modeler. Attendees have the opportunity to create time-series models for historical data, explore and mitigate for seasonality and introduce external factors. Also, see a demonstration of big data forecasting on the desktop, with a hands-on walkthrough of the latest addition, Temporal Causal Modeling, which automatically discovers causal relationships between time series for better forecasts.

Capitalizing on a trial offer

And be sure to to sign up for a complimentary trial of SPSS Modeler. If you are unable to attend IBM Vision 2016, sign up for a complimentary predictive analytics workshop in a city near you. IBM SPSS technical experts will help you get the most value from your 30-day SPSS Modeler trial.

Follow @IBMBigData

This entry was posted in Big Data. Bookmark the permalink.