Optimizely Mobile Proxies for QA of US-Targeted Experiments

Optimizely runs experiments and personalisation that target audiences by location, device and behaviour. An experiment aimed at US mobile visitors in California will never show itself to a QA engineer in Berlin, and a datacenter VPN may be excluded by the audience's own conditions. A Stateside Proxies line in a US metro makes the QA browser a US mobile visitor in that state, so the experiment's variations, the personalisation rules and the goals can be checked as a real visitor triggers them. This page explains the method for QA and growth teams.

What Optimizely audiences read from the visitor

Location from the address: country, state, sometimes city. Device and browser from the user agent and screen. Behaviour from cookies and prior visits. An audience condition such as US mobile visitors in California is met only by a browser whose address is in California and whose device reads as mobile. From the wrong location the experiment stays invisible and the QA pass proves nothing.

Two details shape what a test can show. Location conditions in Optimizely Web Experimentation come from the platform's own lookup of the visitor's address and are not included on every plan, so confirm the state it assigns to the line before assuming a rule has failed. Feature Experimentation, the server-side product, has no such lookup: your application passes in attributes such as state or country, so a wrong result there points at your own code or geo service.

Who checks experiments this way

QA engineers confirming each variation renders correctly for the targeted US audience before an experiment goes live. Growth teams checking that personalisation rules fire for the right states. Analysts confirming that goals and events record for US visitors. Support teams reproducing a US visitor's report of an odd page. Each is the team's own site, visited by the team's own tester from a US connection.

Check both the control and the variation in separate fresh profiles, and record the Optimizely visitor identifier shown in the developer tools alongside each screenshot; a reviewer can then match every screenshot to the results report without guessing which bucket produced it.

Why a carrier line

A real US mobile address in the chosen state meets location and device conditions exactly as a visitor's would, and is not excluded by bot or datacenter filters. The address holds steady, so bucketing stays consistent across a session, and a profile tool keeps the cookies that bucketing relies on.

Setting up the check

Rent a line in a metro in the targeted state, build a browser profile with the metro's time zone and US English, and visit the pages carrying the experiment. Confirm which variation is shown, screenshot it, and complete the goal actions. Check Optimizely's results for the test visitor. Use a fresh profile to be bucketed again, and move the line to another state, free, to check state-specific rules.

Separate design checks from targeting checks. Preview mode and force-variation links can show a variation wherever you happen to be, which is fine for reviewing layout but proves nothing about whether the audience condition works. For that, load the page normally through the line and add optimizely_log=info to the address; the browser console then lists which experiments ran and whether each audience matched, in words a non-developer can follow. Capture that output with the screenshot.

Keeping test traffic clean

Exclude or tag test visitors in Optimizely so results are not distorted, and keep test sessions to the number QA needs. The line exists to see the experiment as US visitors see it.

Optimizely's IP filter can keep named addresses out of experiment results, which suits a fixed address; if you rotate the line, update the filter or your tests will start appearing in results. On a mobile network, also watch for flicker, the brief flash of the original page before the variation paints, and note it when you see it, since slower phone connections expose what office broadband conceals.

Setting up a Optimizely proxy on Stateside Proxies

1. Rent a line in a metro in the state the experiment targets.
2. Create a browser profile with the line's details, the metro's time zone and en-US; confirm the carrier and state on an IP lookup page.
3. Visit the pages carrying the experiment, note the variation and screenshot it.
4. Complete the goal actions and confirm they record in Optimizely.
5. Use a fresh profile to re-bucket, or move the line to another state, free, and repeat.

Optimizely proxy questions