While working at 25sprout as a data analyst, the company faced the challenge of a low sales win rate, despite receiving inquiries from over 10 enterprises every day. Consequently, I built two models to solve two problems.
Firstly, I collaborated with the marketing department to enhance lead quality from marketing toward the sales funnel.
Secondly, I supported the sales department in efficiently allocating operational resources by predicting clients with a higher likelihood of closing deals, thereby improving the overall sales conversion rate.
Prospect will be scored based on their probability to win the deal once entering into the sales pipeline.
(1) Used MySQL to clean, transform data and joined 8 data sources with advanced queries.
(2) Utilized logistic regression to predict the probability of closing the deal and scaling lead-scoring from 0 to 10.
(3) Implemented the lead-scoring model into the pipeline system with API, so the pipeline could score automatically once the prospect led in the pipeline.
(4) Defined metrics, such as Sales Qualified Lead Rate and Marketing Qualified Lead Rate, to track model performance and align marketing investment toward sales revenue, ensuring alignment with company-wide objectives across different departments.
✨ After implementing these two models, the sales cycle decreased by 50% and increased revenue by 180% within 11 months.
✨ Forecasted expected revenue and provided valuable insights for the CEO, heads of sales, marketing, product, and finance department, enabling informed decision-making trade-offs.