How can you be profitable in an oil downturn?
An oilfield equipment provider recently found itself concerned with wellhead price responsiveness during downturns in capital spending, which itself could be attributed to much lower global oil pricing. One of their major goals became building a predictive analytics solution that could incorporate sophisticated machine learning techniques to gain a far greater level of insight and visibility into the probability of winning insight moving forward.
Essentially, they wanted to increase the confidence of each bid through the use of pricing guidance on a per deal basis - which was a goal that the team at Pariveda could help achieve.
Provide the right quote at the right price to predict winning bids
The team at Pariveda worked directly with the company to develop a customized cloud-based machine learning solution that not only used market conditions and customer information, but also comprehensive historical sales data to predict the win/loss probability across the pricing spectrum. This model was then converted into an easily accessible web service and Excel add-in, allowing it to produce a band of "valid" prices and their win probabilities.
Machine Learning navigates complexity in oil industry
Thanks to the data-driven pricing decisions that are now available during the quote preparation process, the company's bidding team can now see what effect price changes have on win/loss probability for an individual deal in real-time. This bold new approach not only guides pricing, but also better informs sales and speeds up the quoting process at the exact same time. The helpful Excel interface also provides much better insight for the marketing department with a mechanism that tests various system configurations, market conditions, and customer segments.
Jupyter Notebooks powered by Python 3 are also used to extract and transform the SAP data, feeding the predictive Two-Class Boosted Decision Tree model in a way that is easy to access and even easier to depend on moving forward.