Implementing Machine Learning in Account Based Marketing



A large technology company had a limited understanding of which commercial customers were most likely to buy its products leading to untargeted, inefficient sales and marketing efforts. An account-based marketing approach was being pursued to help improve cross-sell on over $50 billion of sales annually, however insight was limited to direct the sales and marketing activities. This limited understanding left untapped revenue lift opportunity.




To determine which accounts should be prioritized for sales and marketing activity, a machine learning algorithm was created to determine accounts with the highest probability to convert to a sale. Unify developed a custom machine learning model using Python. The solution predicted each individual commercial customer’s probability to convert, as well as the automation to surface these results with relevant information to help with messaging and positioning. Each customer was assigned a probability, and a threshold was created to ensure only the accounts that were most likely to be purchased were recommended to sellers.

Key Outcomes


Customer accounts that are recommended are expected to convert at a rate of  two times better than random guessing. The direct result is increased revenue and efficiency in the sales and marketing field. Typical revenue lift for similar programs is 10 to 30%.