When your internal dashboards show green across every service-level objective, it is natural to assume the customer experience is healthy. For a global scientific instruments company serving researchers, laboratories, and distributors in 28 countries through a multi-cloud digital platform, that assumption turned out to be wrong.
We recently published a case study examining what happened when this company introduced outside-in monitoring running from over 110 nodes to measure the experience customers actually had. The findings reshaped how engineering, operations, and vendor teams understood their digital customer journey.
The gap between SLOs and customer experience
Availability exceeded 99.5% across all five monitored endpoints. Every cloud provider cleared every SLO. By any internal measure, the platform was performing well.
But API monitoring from the customer's perspective told a different story. Core markets in North America and Europe loaded pages in roughly 210 milliseconds. Other markets were nearly three times slower.
A home-market view of the dashboards would have declared the estate healthy.
Where the time actually goes
The latency breakdown exposed two distinct performance profiles that internal monitoring had not separated.
On transaction pages — cart creation, cart retrieval, cart deletion — backend processing consumed 84–86% of total load time. DNS and network handshakes were fast and consistent; the bottleneck was application-layer work.
On catalog and search pages, the pattern reversed. DNS resolution accounted for a surprisingly large share of total time, making edge and DNS optimization the primary lever.
Before this breakdown, teams had assumed network paths were the dominant issue everywhere. The per-component, per-endpoint evidence redirected optimization effort to the right layer for each page type.
What changed as a result
The outside-in baseline produced four concrete operational outcomes:
-
An edge optimization roadmap. Regional page load data quantified the growth-market gap. The nearly 3× difference between core and underserved markets is an infrastructure decision, not a physics constraint.
-
Backend tuning priorities. The latency breakdown pinpointed transaction pages as the optimization target — and showed that catalog pages would benefit most from DNS and edge tuning.
-
Vendor accountability. Cloud provider comparisons moved from anecdote to data. Median, p75, and p95 analysis gave procurement and operations teams a factual basis for vendor conversations.
-
Persistent assurance. A continuous baseline means new deployments, edge policy changes, and cloud provider routing updates are measured against the same benchmark — not re-evaluated from scratch each time.
The pattern that keeps recurring
This case study illustrates a pattern we see repeatedly across engagements: organizations with mature monitoring that nonetheless lack an independent, outside-in view of the customer experience. Internal dashboards answer "is the infrastructure up?" Outside-in monitoring answers "what does the customer actually experience?" — and those are different questions with different answers.
Download the full case study to see the complete latency breakdown, cloud provider comparison, and regional performance analysis.