The commercial area of the company in which the project was developed faced some problems to adjust their prices because they held more than 20,000 SKUs. Consequently, many clients gave up on their contracts, due to price divergences if compared to competitors, or also because they were not able to offer a price more compatible with the intended value delivery. Another challenge the company had to face was a logistics price increase, due to the container crisis.
That volume of SKUs, and the dynamics affecting costs due to the crisis, forced price adjustments to be made rather quickly and assertively; however, the difficulties posed by that process led to analysis times and adjustments of up to 6 months.
In clearer terms, the commercial team’s needs involved:
- Establishing a price for all products, in a individualized and dynamic way, by taking into consideration their technical characteristics, the profile of each client at the global level, and the different sales regions and exchange rates (forex);
- Guiding each sales segment with information on potential products to increase their net revenue, in comparative terms;
- Providing an average price history;
- Providing the products’ Weighted Average Price per Volume (VWAP);
- Having an efficient tool to negotiate sales, and so enable them to maximize their profit margin during negotiations.
By using the CRISP-DM methodology, all the initial understanding was built during the first stage, of Business Understanding, where the effort was concentrated on understanding the commercial process, the opportunities, gaps, pains and needs, and the impact of that process on the company’s value chain. For this process, the KBQs (Key Business Questions) were used to guide the entire project.
That understanding orientated the creation of an AI solution to provide, in a dynamic way, the prices for the SKUs. The macro plan for that solution considered the following items:
- Generating a set of regression linear models for product prices in the whole market where that company operates, including the possibility of performing analyses at the local level.
- Expansion of the technical information on the materials, including hybrid products (single products, non-detachable and designed to perform the function of two or more products) in the analysis.
- As much information about the client as possible to distinguish its area of operation.
- Monitoring of the average ticket evolution to avoid any hustle.
- Looking for opportunities to increase net revenue in the largest number of possible scenarios, and in a consistent and comfortable manner.
Examples of parameters contemplated by the algorithm are: intrinsic information on the material (capacity, voltage, among others), as well as parameters from the markets in which those materials are being sold, and information about the clients and their areas of operation.
The AI is based on a price list and, through an API that follows exchange rate variations in an integrated manner with the solution, it obtains the exponential moving average in the last 30 days in the foreign exchange market.
The developed solution was a web application that enables users to compare practiced prices, for any material, with the prices suggested by the AI algorithm in a matter of seconds. The solution also includes a range of filters and parameters, such as sales region, family of products, and client profile, among others, to help users direct their analysis if so desired. A user is also able to view the impact any exchange rate fluctuation might have had on prices during the selected period, thus making the solution even more robust.
The pricing analysis based on the AI algorithm has allowed the company to expand its analysis capacity much beyond what would be possible for a human being to perform, by simultaneously contemplating dozens of parameters, and has enabled the company to make data-driven decisions. As a consequence, the solution has been able to provide a return of up to 4% over the expected revenue when practicing prices based on the AI algorithm.
At the end 2022, Nidec Global Appliance’s IT team was awarded by 4Network – IT4CIO Portal with the NIDEC Lakehouse case of support to Intelligent Pricing.
Testimonies about the case:
“At the end of 2022, Nidec Global Appliance’ IT team was awarded by 4Network – IT4CIO Portal with the NIDEC Lakehouse case of support to Intelligent Pricing.
That delivery was the result of a joint work undertaken with Dojo Smart Ways, a Nidec partner company specialized in data processing.
The data-driven journey generates an increase in analytical capacity so that managers can focus less on the consolidation process and much more on making assertive decisions.
This project led by me has eventually healed the company’s pains: how to respond to an increasingly challenging pricing scenario.
With their global reach, the Embraco / Nidec Global Appliance products count on capillarity among many clients and a vast range of SKUs to fulfill any and every need in sectors such as refrigeration, medical applications, food preservation, and vending machines, among others. In such a complex scenario, the use of technology is fundamental to guide the company in its mission to optimize results through a sound price policy.
At NIDEC Data Lakehouse (our data platform) we gather and transform all the databases necessary to perform price analyses (and there is not just a few of them, lol). With everything being ready and consolidated in what we call a “trusted” layer, a web portal consumes that information and makes data available to be visualized. Here users are able to use many different filters to perform their customized analyses and look for new insights.
We also employ an algorithm in this same solution that, by using historical data and specific conditions entered by the user, returns a price recommendation that orientates the work of managers”.
Matheus Dantas de Queiroz
