Technology and infrastructure
Built in house, because the numbers depend on it.
Owning the stack is not a preference. It is what lets us attribute a session to a partner, price inventory on evidence, and answer a due-diligence question without asking a vendor first.
- Analytics
- First party
- Attribution
- Server side
- Delivery
- Multi-provider
- Third-party tags
- None
The stack
What we operate ourselves.
High-volume web infrastructure
Web serving, caching and database layers tuned for sustained high request volume rather than for a demonstration.
First-party analytics
Our own event collection and a columnar analytics store. There is no third-party analytics tag on our properties, which keeps the data ours and the privacy position simple.
Traffic attribution
Resolved server side at the landing event and carried through the session, so a partner, campaign and source stay attached to an outcome rather than being inferred later.
Content delivery
Delivery across multiple providers in front of origins we control, with routing and cache policy managed per region.
Data processing
Aggregation runs as scheduled jobs over rolled-up tables rather than over raw events, so reporting stays fast as volume grows.
AI applications
Machine learning applied inside our own platforms, and consumer-facing AI products developed on infrastructure we own and operate.
Subscription technology
Billing, entitlement, account and reconciliation systems behind our subscription products, including payment provider integration.
Privacy products
Consumer privacy technology built around minimal collection, including our own VPN service and its routing layer.
Data handling
We collect the minimum the job needs.
Our analytics exist to answer commercial questions about traffic and monetisation, not to build profiles of individuals. Events are recorded against a session identifier rather than an identified person, and reporting is aggregate.
Full IP addresses are not retained beyond the point where they are operationally needed. Geography and device are resolved at collection time and only the resolved classification is stored. Reporting uses internal property aliases in place of property names wherever the name is not required.
What we hold, why we hold it and how long it stays is set out in the privacy policy.