Independer uses AI to predict conversion value and optimize marketing performance
Conversion probability model enables smarter targeting, bidding, and media efficiency
Independer is one of the leading comparison platforms in the Netherlands, helping millions of consumers make informed choices in complex markets such as insurance, energy, and mortgages. Within the highly competitive market for health insurance, performance marketing plays a crucial role in reaching the right users at the right moment.
To stay ahead, Independer is moving beyond descriptive analytics and uses data to actively predict which visitors were most likely to convert, and adjust marketing efforts accordingly.
The challenge
Independer attracts a large volume of traffic to its platform, but not every visitor has the same likelihood of completing a health insurance deal. Traditionally, marketing optimization relied on historical conversions and aggregate performance metrics. This approach had clear limitations:
- Media spend was distributed relatively evenly, regardless of a user’s true conversion potential.
- Retargeting campaigns included all different types of users, including those who were either very unlikely to convert or already almost certain to do so.
- Bidding strategies were not aligned with the expected value of individual users.
Independer needed a way to predict future conversion behavior, rather than only analyze past results, and use those predictions directly in its marketing platforms.
The solution
Together with Relevant Online, Independer developed a conversion probability model using Google Analytics 4 data in BigQuery, powered by Google Vertex AI (AutoML).
The model predicts, for each individual user, the probability of completing a health insurance deal on Independer.nl. Based on this prediction, users are assigned a probability score and segmented into low, medium, or high conversion likelihood groups.
Key characteristics of the solution:
- Built entirely within Independer’s own Google Cloud environment, ensuring full data ownership and control
- Trained on two years of historical GA4 data to capture behavioral patterns leading up to conversion
- Outputs a probability score per user that can be activated directly in marketing channels
- Designed to support real-time audience creation, bidding strategies, and exclusions
The key of AI-driven prediction is its specificity. This setup enables Independer’s performance marketing team to use AI-driven predictions as a core input for campaign optimization.
Impact
By integrating conversion probability into its marketing strategy, Independer unlocked several high-impact use cases:
-
Smarter prospecting
High-probability users are used as an audience for lookalike targeting in platforms like DV360 and Google Ads. This improves the quality of incoming traffic by focusing prospecting efforts on users who resemble the strongest converters. -
More efficient retargeting
Users with a very low probability of converting are excluded from retargeting campaigns, reducing wasted spend. At the same time, users with a very high probability can also be excluded, preventing overinvestment in users who are likely to convert anyway. -
Value-based bidding
Bids are adjusted based on predicted conversion likelihood. Instead of a one-size-fits-all approach, media spend is allocated according to expected value, leading to more efficient budget use and improved cost per lead.
Result
With this AI-driven approach, Independer is one step closer to a fully specific predictive, value-based marketing strategy. Campaigns are now optimized based on who is most likely to convert, not just who converted in the past.
The result is a more scalable, efficient, and future-proof performance marketing setup, where data and AI play a central role in driving smarter decisions and better outcomes.
Learn more