Inconsistent Geolocation Results Across Multiple IP Geolocation APIs: How to Reconcile Discrepancies?

Following up on the previous discussion about IP lookup accuracy, I'm encountering significant inconsistencies when using different IP Geolocation APIs.

My application relies on precise geolocation for regional content delivery and analytics. I've noticed that for the same IP address, especially mobile or VPN IPs, various IP Geolocation APIs return widely different geographic locations (e.g., different cities, states, or even countries). This directly impacts user experience and location data reliability, which is crucial for our service.

Here's what I've tried to address this:

• Integrated and tested with several prominent IP Geolocation APIs (e.g., MaxMind GeoIP2, ipinfo.io, Abstract API).
• Created a test suite with a diverse range of IP addresses, including known residential, mobile network, and VPN IPs.
• Implemented basic retry logic and error handling for API calls.
• Verified network connectivity and firewall rules on my end to ensure no interference.

Despite these efforts, the discrepancies persist. For a particular mobile IP, API A might return 'New York, NY', while API B returns 'Philadelphia, PA'. The expected location is often somewhere in between, or a completely different region altogether, making it hard to trust any single source. Here's a simplified console output illustrating the issue:

API A Response: {"ip":"192.0.2.1","city":"New York","region":"NY","country":"US"}
API B Response: {"ip":"192.0.2.1","city":"Philadelphia","region":"PA","country":"US"}

Given these challenges, I have a few questions for the community:

• Is this level of discrepancy common among IP Geolocation APIs, or am I missing a crucial configuration?
• Are there established techniques or services for reconciling conflicting geolocation data from multiple providers to achieve higher confidence?
• Can anyone recommend an IP Geolocation API that tends to be more consistent and accurate for mobile and VPN IPs?
• What are the best practices for handling scenarios where geolocation accuracy is paramount, given these inconsistencies?

Thanks in advance for any insights or recommendations!