SETTING THE BEST PRICE

The products sold by the manufacturer of equipment are both very complex and highly customizable. On the surface, they only have about a dozen or so main products, but each come in different sizes and configurations and they may be built with any of several hundred different add-ons that also come in different sizes and specifications. In total, there are thousands of different product combinations to consider.

​

In the past, their standard pricing had mainly been based on the cost of manufacturing each component taking into consideration the fluctuating market prices for some of the expensive raw materials that were used in the manufacturing process. The account manager, who would prepare the quote based on the specifications from the customer, would then adjust the final price as they considered appropriate to win the contract while still making a profit. The first part of setting the price was very sophisticated, complex, and data-driven; the last part was based purely on intuition and the limited past experience of the individual agent.

​

I worked with the pricing analytics team to help them usher in a new methodology for their consumer pricing that was based on data from previous pricing quotes that were won or lost rather than just costs.

They had realized that their customers willingness to pay for the final product was not directly related to the cost of manufacturing and that the sum of the whole was not necessarily equal to the sum of its parts. The new approach relied on the data that had been collected from winning bids and the information that was available for bids that were lost, which often included the reason that the bid was lost and which competitor won the bid.

​

The pricing team worked under my supervision and mentoring to gather, cleanse, and prepare the data on the historical bids including the detailed configuration of each product down the size and color of each nut and bolt as it had been quoted. This was combined with information about the customers and the outcome of each bid as described above.

​

All of this data was then parsed through a series of different models ranging from simple linear regression over decision trees to neural network models to determine the winning price for each product configuration. One of the important new insights from this process was the impact of geographic location on prices, which the company discovered was an important differentiator.

​

Having the internal pricing analytics team (as opposed to external consultants) build this new approach and take full ownership of the results was extremely helpful to the adoption of the new pricing methodology. They already have credibility and the trust of the account managers and having worked on the details of these models, they can talk confidently about how they arrived at a price recommendation for a given configuration and they are able to quickly make corrections to the data preparation or the models based on the feedback from the account managers.