Two Offices, One Disagreement

[brownnadean]

Our London and Singapore teams are reviewing the same public New Zealand pages, but they keep reporting different loading times. Both use proxies set to New Zealand, so management expected the numbers to match. I suspect we are mixing the speed of the target site with the distance and routing between each office and the proxy service. How would you design a fair comparison? We need a report that says what was measured and where, instead of declaring the site slow based on two people pressing refresh from different networks at different times.

[nessie71]

I would split the question into two tests. One measures the full journey from each office through the proxy, which is relevant to those teams’ workflow. The other uses a fixed test runner and consistent conditions to compare changes in the target pages over time. Keep the requested pages, browser version, cache state, and schedule aligned, then report a range across repeated runs. A New Zealand exit IP identifies one part of the route; it does not make the whole path identical. Label the results with the test origin so the report does not present office network differences as a property of the target site.

[879lilia]

That distinction would prevent a lot of arguments over a single average. I would state the question above each chart and list the test origin beside it. Keep office productivity measurements separate from comparisons made by the fixed runner. The page at simplynode.io/locations/new-zealand-proxies offers plenty of useful information about New Zealand proxy connections and session options. Use it to document the routing setup, then measure timing with your own repeatable test. Run the same page set at planned intervals and retain the slow results as well as the fast ones. The final report should explain the tested route instead of implying that every visitor in New Zealand sees the same speed.