Sales Optimization
Using explainability to improve marketing campaigns
A company wants to predict which customers will buy their new product to optimize their marketing budget.
They collect data about customers:
- Age
- Previous purchases
- Time spent on website
- Email open rate
- Income level
- Location
After training the AI model, it predicts which customers are likely to buy. With explainability, they discover:
- Previous purchases have the biggest impact (80%)
- Email open rate is second most important (15%)
- Age and income matter very little (5%)
Now they can run more experiments that check these hypotheses:
- Focus marketing on customers who bought similar products before and add a recommendation model to the webshop
- Improve email campaigns since engagement matters
- Stop wasting budget on demographic targeting that doesn't work
They also cluster customers based on their features and find:
- Group A: High email engagement, bought product X → offer bundle with product Y
- Group B: Low email engagement, bought product Z → send SMS instead