Using Analyze With Localize and Optimize

CAS.org runs localized content, and pages can also run experiments through Webflow Optimize. Analyze handles both, with one behavior that surprises people.

Click Events Are Combined Across Variations

When a page has localized versions or optimized variations, Analyze shows click events for all variations together. The base version, the localized versions, and any optimized variations report as one aggregate figure per event.

This means a top event count on a localized page is a site-wide total for that element, not a figure for one language or one test variation.

What to Do About It

• Treat event totals as global. Do not present an event count as evidence about one locale.
• Use page level metrics for locale questions. Sessions, bounce rate, and time on page are more reliable than event counts when comparing locales.
• Use Optimize for test results. When an experiment is running, the experiment reporting is the source for which variation won. Analyze describes traffic and engagement around it.
• Say which figure is which. When a report mixes aggregate event counts with locale-specific page metrics, label them so nobody reads one as the other.

Localized Pages

Localized pages are tracked like any other published page on an allowed domain. For how localization is structured on CAS.org, see Translation and Localization.

Two things to keep in mind:

• Traffic volumes differ by orders of magnitude between locales. A locale with low traffic will produce percentages that swing wildly. Widen the date range before drawing conclusions.
• Translated button copy changes labels. Because Analyze builds labels from element text, translated wording affects how events appear. See Naming Elements So Click Events Are Readable.

Pages Running an Experiment

• Do not change the page mid-test. Editing content while an experiment runs makes the result unusable.
• Check goals before launch. A goal created after the test starts has no data from the beginning of it.
• Give the test a full window. Ending early on an encouraging first week is the most common way to reach a wrong conclusion.