Maintaining a competitive pricing strategy using Machine learning model
E Energy receives a file from the regulator every two weeks with all the gas and electricity tariffs for competitors in the market. A relatively large administration team would then embark on deciphering the huge volume of data to help understand where E-Energy’s price sat in comparison to others in the marketplace to remain competitive.
Administration requirements were so large because the format of the file and the data within fields was not consistent. It had anomalies and variations from month to month and to be useful, the team had to cleanse the data, remove duplicates, correct formatting and then run manual scripts to check the data quality of the file. This process can take up to 3 working days, with significant senior resource also involved required.
Understanding whether Machine Learning was the right technology
We mapped out the irregularities of the model and identified that cleaning and reformatting of the tariff file was essential automation. From this, we developed an automated machine learning model that enabled E-energy to understand the competitor market.
Our developers trained our model on existing data sets and now it removes the duplicates and fixes the values within the file and highlights any new anomalies in the data. The output file is 100% accurate and it only takes 55 seconds for the whole process from start to finish, providing significant resource and time efficiencies.
ML created a competitive edge and an annual saving to the business of over £30,000 per annum
