These are two services we run ourselves and offer to clients. They exist because we needed them in production, kept hitting the same operational problems, and ended up building the platform we wished was already there.
Landing pages hosting ¶
Hosting platform for landing pages, optimised for fast time-to-first-byte and predictable behaviour under traffic spikes. Pages are served from POPs across North America, Europe, and Asia, with HTTP caching and a small edge runtime for redirects and A/B logic.
What it does:
- Edge serving. Static assets and pre-rendered HTML go out from the closest POP. Cache rules are explicit per route, not magic.
- Failover. If a POP is degraded, traffic shifts to the next one without a manual page-out. We test this on a schedule, not just when something breaks.
- Burst capacity. The platform is sized for steady traffic and scales horizontally for campaign launches. We size headroom against measured peaks, not optimistic projections.
- Hardening. TLS everywhere, regular dependency and infrastructure scans, basic WAF rules, and an incident playbook the team has actually rehearsed.
- Visibility. Per-page analytics: requests, latency percentiles, cache hit ratio, error rates. Same dashboards we use to operate it.
What it is not: a generic CMS, an enterprise CDN replacement, or a magic conversion booster. It is a focused hosting service for landing pages, run by people who understand both the operational and marketing side.
Traffic distribution system ¶
A traffic distribution system (TDS) for routing visitors to the right offer or destination. The rule engine supports more than 50 conditions: geography, device, OS, language, time of day, referrer, custom user attributes, and combinations of all of those.
Notable bits:
- Rule engine. Rules are composable. A common pattern is a country filter, then a device split, then a campaign-specific rotation. Decisions are evaluated at the edge, not on a central server.
- Performance. Routing decisions are O(1) per request after rule compilation. P99 added latency stays in the low milliseconds even under load.
- Targeting. Routes can be A/B tested, weighted, capped per click or per period, and switched off without redeploying anything.
- Analytics. Click-level data exported to your warehouse of choice (BigQuery, Snowflake, ClickHouse) plus in-product dashboards for day-to-day work.
- Reliability. Multi-region active-active. Loss of one region is a metric on a graph, not a customer-facing outage.
- Security and privacy. TLS, signed click identifiers, configurable PII handling, and retention policies you can actually adjust.
The TDS runs from POPs on multiple continents, so traffic is decided close to the user rather than round-tripped to a single origin.
If either of these matches what you need, get in touch. Happy to walk through the architecture and where it does and does not fit.