Geolocation Data, Explained: Understanding Signals, Accuracy, and Real-World Impact  

In reality, IP geolocation is one of the most complex internet data challenges – because the internet itself constantly moves. Networks expand into new regions, prefixes get reassigned, infrastructure shifts between data centres, and traffic can exit from locations that look legitimate on paper but aren’t physically real. 

To explore what accuracy really means in this context, IPXO hosted a webinar with Oliver Gasser, Head of Research at IPinfo, moderated by IPXO Co-Founder Vincentas Grinius. The discussion unpacked how modern geolocation datasets are built, what signals matter most, and why evidence-based validation is essential for keeping data trustworthy. 

Geolocation Is Not One Dataset – It’s a System of Signals 

A key takeaway from the webinar is that geolocation isn’t “one database that knows everything.” Instead, it’s an aggregation problem: collect many signals, rank them, validate them, and continuously adjust. 

These signals (or “hints”) can come from multiple layers of the internet stack and operational ecosystem, including: 

• GeoFeeds (standardised location hints published by network operators) 
• WHOIS attributes (often country-level, sometimes more granular) 
• reverse DNS patterns (sometimes encoding location clues like airport codes) 
• device and network observations (feedback loops from apps and partner sources) 
• active measurements (latency tests, traceroutes, port scans, etc.) 

The challenge is that some signals are clean and intentional (like GeoFeeds), while others are indirect and messy (like reverse DNS). Some are updated regularly. Others can stay wrong for years. 

This is why geolocation providers don’t just collect data – they need an internal engine that can decide what to trust. 

GeoFeeds – The Best Standard We Have (But Not a Perfect One) 

GeoFeeds are one of the most important developments in IP geolocation quality. They provide a structured, standardised way for network operators to publish prefix-level geolocation information in a CSV format, aligned with IETF standards. 

In theory, GeoFeeds should be one of the strongest hints because they come directly from operators – the people who know where infrastructure is actually deployed. 

In practice, GeoFeeds still come with issues: 

• incorrect entries 
• formatting and parsing problems 
• outdated information 
• mismatched prefix granularity 
• inconsistent publishing practices 

The webinar also touched on how the IETF has been involved in this area, including the standardisation of GeoFeeds and community discussion around how to improve the ecosystem. 

GeoFeeds help, but they don’t eliminate the need for measurement. They simply raise the baseline quality – if providers treat them carefully. 

Why Validation Matters – Latency Can Prove What’s Impossible

The most valuable part of geolocation isn’t the hint collection. It’s the validation layer. 

One of the strongest validation approaches is latency measurement. If a provider measures round-trip time from multiple global vantage points, they can estimate whether a claimed location is even physically possible – using speed-of-light constraints (and the fact that fibre optics slow signals down further). 

This is where geolocation becomes evidence-based instead of assumption-based. 

That principle is powerful because it doesn’t rely on “trust.” It relies on physics. 

The Role of Probe Networks – Coverage Determines Confidence

Measurement quality depends on where you can measure from. 

Oliver explained that IPinfo runs a global probe network with: 

• 1,300+ points of presence 

• 140+ countries 

• 540+ cities 

This matters because geolocation accuracy is highly dependent on geographic diversity. It’s not enough to measure from Europe and North America and assume the rest of the world behaves similarly. Regions with limited infrastructure, long routing paths, or sparse measurement coverage are exactly where false assumptions are most likely. 

The discussion also compared different measurement platforms. Tools like RIPE Atlas provide strong community-driven coverage, but running an internal probe network offers flexibility – enabling custom measurement types at scale. 

VPNs and Virtual Locations – When Geolocation Becomes a Marketing Layer

One of the most practical sections of the webinar focused on VPNs. 

VPN providers often claim presence in a large number of countries, but these claims aren’t always true. Some “locations” exist only on a website – not as physical exit nodes. 

IPinfo’s research found that: 

• most providers had cases where traffic exited a different country than claimed 
• some countries had no physical exit at all, despite being advertised 
• virtual locations tend to appear in regions where deployment is costly or difficult 

This is not a small issue. It affects: 

• fraud prevention systems 
• compliance workflows 
• licensing decisions 
• user security assumptions 
• any business logic built on “country of origin” 

In other words, inaccurate geolocation isn’t just annoying – it can create legal and operational risk. 

Update Frequency – Daily vs Monthly Makes a Real Difference

Another point that deserves more attention is update cadence. 

If geolocation changes frequently, slow updates produce drift – and drift becomes wrong data. 

IPinfo’s approach includes: 

• daily database updates 
• real-time freshness via API access 

That matters because the internet changes continuously. Prefixes move. Traffic patterns shift. Infrastructure expands. Leasing changes usage. And VPN exit nodes rotate. 

If a company refreshes their dataset monthly, they can be working with data that’s already outdated by the time it’s deployed. 

Why Geolocation Changes So Often

A surprising insight from the webinar is just how much churn exists in location mapping. 

Oliver mentioned that geolocation databases can see 10%+ monthly city-level changes. That’s huge – and it reflects the reality that geolocation is not a static mapping. 

Drivers include: 

• IP leasing and reassignment 
• infrastructure migration between regions 
• multinational networks shifting capacity 
• CDN expansion 
• traffic engineering decisions 
• anycast deployments 

This also explains why “accuracy” is difficult to define without context. A location can be correct today and wrong tomorrow – not because the database is bad, but because the network moved. 

What This Means for IPXO and the Leasing Ecosystem 

For IPXO, this topic is directly relevant. 

Leased IPv4 space is inherently dynamic. Prefixes move between tenants, regions, and use cases. This affects geolocation, reputation, abuse monitoring, and customer trust. 

The webinar raised an important point: when an IP range is parked (not actively used), its geolocation should not automatically inherit a previous tenant’s location. That approach reduces false assumptions and makes it easier for downstream systems to interpret location responsibly. 

Ultimately, geolocation quality depends on transparency, measurement, and speed – which aligns with IPXO’s broader mission to bring clarity and structure into IP resource usage. 

FAQ