Our customer built models to predict top-of-funnel and initial form submission conversion rates, but they were producing too many inaccurate predictions. The company was having trouble pinning down the factors that indicated a higher likelihood of conversion based on the data collected from their own website. They could gather data such as click locations and pages visited, but they couldn’t build a full picture of the visitors.
The company built two parallel models — top-of-funnel and initial form submission — using Explorium to identify the right factors for conversion rates. Our customer connected their datasets to the Explorium Enrichment Catalog to add the following data to each model:
Start of funnel conversion prediction:
Form submission conversion prediction:
The company’s top-of-funnel model saw an immediate uplift in predictive accuracy to 88.6%, leading to a much higher conversion rate and allowing for smarter marketing spend upfront. The initial form submission model had a 90.44% accuracy score when predicting conversions. Combined, these two models allowed the company to make smarter marketing decisions and to understand which actions would lead to user conversion. As a result, our customer was able to design a better user journey, and boost its revenues and customer engagement significantly.