What's the best way to implement geolocation for GEO-targeted content? IP lookup vs Geolocation API debate

We’ve been running geo-targeted content at scale for 5 years. Here’s what works:

Provider recommendation:

MaxMind GeoIP2 for most use cases. Why:

• 99.8% accuracy at country level
• ~80% accuracy at city level (varies by region)
• Good update frequency (weekly for paid)
• Solid API and local database options

Our architecture:

1. Edge CDN detects IP -> country/region (Cloudflare Workers)
2. Initial page load uses IP-based location
3. If precise location needed, prompt for Geolocation API
4. Cache location preference in cookie for return visits

On VPNs:

You can’t fully solve VPN detection without getting creepy. Our approach:

• Detect obvious VPN/proxy IPs (MaxMind has flags for this)
• Show a “Confirm your location” option when we suspect VPN
• Let users manually override their detected location

Accept that ~5-10% of users will have incorrect location detection. Build your UX to handle that gracefully.

I work on geolocation systems. Some nuances to consider:

Provider comparison:

The accuracy reality:

• Country: All providers are 99%+
• State/Region: 85-95%
• City: 65-85% (highly variable)
• ZIP/Postal: 50-70% (not reliable)

My recommendation:

For your e-commerce use case, MaxMind is the sweet spot. If you need city-level accuracy for critical features (like showing local stores), combine with optional Geolocation API.

Don’t promise users city-level accuracy with IP alone - you’ll disappoint them.

Privacy engineering perspective here. Consider GDPR implications:

IP-to-Location:

• IP addresses are PII under GDPR
• You need legal basis for processing
• Legitimate interest usually applies for geo-targeting
• Document your approach in privacy policy

Geolocation API:

• Requires explicit consent
• More privacy-friendly (user chooses to share)
• But adds friction to UX

Best practice:

Use IP for functional purposes (pricing, availability). Use Geolocation API only when there’s clear user benefit (store locator, delivery estimation).

Don’t collect more precise location than you need.

Mobile-specific considerations:

Geolocation API on mobile:

• Much more accurate (GPS available)
• But requires explicit permission dialog
• Permission fatigue is real - users often deny
• iOS especially strict about location access

Our mobile approach:

1. IP-based on initial load (no permission)
2. Only request GPS for specific features (store finder, delivery)
3. Explain WHY before the permission prompt
4. Have a clear fallback for denied permissions

The stats:

When we ask for location with context (“To show nearby stores”):

• 65% grant permission

When we ask without context:

• 30% grant permission

The explanation matters a lot.

Caching considerations for geo-targeted content:

The problem: Page caching + geo-targeting = users seeing wrong content

Solutions:

1. Vary header approach:

   • Vary: CF-IPCountry (or similar)
   • Creates separate cache entries per country
   • Can explode cache size if too granular

2. Edge compute approach:

   • Run geo logic at CDN edge
   • Inject location data before page render
   • More flexible, handles personalization better

3. Client-side approach:

   • Cache generic page
   • Fetch location-specific content via AJAX
   • Simplest caching, but content shift on load

Our setup:

Edge compute for critical geo content (pricing, availability). Client-side for nice-to-have personalization.

Don’t try to cache city-level personalization - the cache hit rate tanks.

For anyone on WordPress, there are plugins that handle this:

Recommended plugins:

• GeoTargetingWP - Solid IP-based targeting
• If-So - Good for conditional content
• WPEngine Geolocation - If you’re on WPEngine hosting

Our experience:

We use GeoTargetingWP + WP Rocket (caching).

Key settings:

• Exclude geo-targeted pages from cache
• Or use separate cache entries per country

The plugins handle the IP lookup, you just configure the rules.

For custom development, they can be limiting. But for content personalization, they work well enough.

One pattern I see work well: progressive enhancement.

The flow:

1. Immediate (0ms): Use CDN’s built-in geo detection (Cloudflare cf-ipcountry, AWS CloudFront-Viewer-Country)
2. Fast (50-100ms): Enhance with IP2Location/MaxMind for city-level
3. User-triggered: Only use Geolocation API when user takes action requiring precise location

Example implementation:

// On page load - country from CDN header (free, instant)
const country = getCDNCountry();

// For enhanced features - IP lookup (fast)
const city = await getIPCity();

// Only when needed - browser GPS (user permission)
const precise = await getPreciseLocation();

This balances speed, accuracy, and user experience.

Great discussion. Here’s our implementation plan:

Architecture:

1. Cloudflare Workers for country detection at edge (free, instant)
2. MaxMind GeoIP2 for city-level when needed
3. Geolocation API only for store finder feature
4. Cookie-based preference storage for overrides

Key decisions:

• Country-level: Edge CDN (cf-ipcountry)
• City-level: Server-side MaxMind (cached)
• Precise location: Opt-in Geolocation API
• VPN users: Manual location override option

Caching strategy:

• Separate cache entries by country
• City-specific content via AJAX (no caching issues)
• Location preference in cookie (survives cache)

Privacy:

• Document IP processing in privacy policy
• Only request GPS when there’s clear user benefit
• Allow location override for all users

Thanks everyone for the practical insights. The hybrid approach is definitely worth the complexity for our use case.