It may be that big data, rather than crystal ball guesses or gazing into the past, is the best way to make the most important decisions. However, it doesn't always happen.
At a recent conference, I had a fascinating conversation with a senior director at a semiconductor manufacturer. She was responsible for shaping her company's pricing management strategy, and she was in the midst of leveraging technology investments made over several years to craft and execute company strategy.
The company believed that, in a large percentage of cases, its pricing decisions were less than optimal, and it was not realizing the best price. In absence of information that would have allowed the company to manage pricing at a product/customer/ microsegment level, the sales team generally relied on history to make pricing decisions.
Over the past few years, this scenario has been common across the high-tech value chain. Facing complex bills of materials containing several thousand unique components and subassemblies, and with limited data and automation, product marketers have increasingly depended on factors such as cost, volume, historical pricing, and past experience to formulate pricing strategy.
Leveraging additional data points such as availability, competitive pricing and alternatives, approved vendor lists, currency conversion, duties and taxes, and engagement terms and conditions at a product/customer level has been difficult at best. Either the intelligence simply did not exist in a format conducive to analysis, or the analysis was not available in near real-time. In absence of useful intelligence, product and sales people have frequently responded to competitive pressure by offering discounts, rather than risk losing the business.
The drive to optimize pricing is not new for most companies, but the shift toward a new generation of analytics has created opportunities and some momentum. Capabilities can now go beyond the analytics centered on business intelligence (BI). Econometrics, third-party subscriptions, social and web-centric data, and sensor data can add significant intelligence.
With the arrival of big data, the rise of predictive analytics, technology to support near real-time modeling and visualization, and a large population of algorithms, it is increasingly possible to address the bottlenecks cited in the preceding paragraph. As a result, not only is it feasible to analyze a multitude of data points to arrive at pricing decisions. It is also increasingly feasible to deploy predictive analytics to understand the effectiveness of those pricing decisions. As data points change, predictive analytics can add context in real-time to pricing decisions and arm product and sales people to negotiate an optimal price.
Some industries (such as retail) have been at the forefront of analyzing pricing elasticity down to an individual product level, and they have leveraged the intelligence to manage pricing decisions in real-time. Nevertheless, pricing management has been slow to take off in B2B engagements. Some members of the high-tech supply chain are finally taking note. Companies that have created robust traditional BI systems and an effective customer segmentation strategy based on purchasing behaviors may be best positioned to move to the next level.
The senior director who talked to me said her goal was to address four challenges as a part of her company's pricing management initiative:
- Identify underpricing at a product/customer/microsegment level.
- Predict the impact of pricing changes at the product/customer level, so such changes can be implemented where feasible.
- Develop insight into additional products that could be sold profitably to current customers.
- Optimize product availability and pricing to maximize revenue.
I invite you to share your insight and experiences related to the application of predictive analytics in B2B environments for pricing management and optimization.