Metrics

Metrics are how A vs B knows whether your experiment succeeded or failed. You define what a "conversion" means: a button click, a page visit, a form submission, a purchase. The platform then measures how often each variation (one specific version being tested, the control or one of the challengers) achieves it.

This section covers the everyday metric types first: clicks, pageviews, custom events, and revenue. Further down are the advanced types built for expert users. These are ratio metrics, quantile metrics, composite metrics, and winsorization (capping extreme values, like one giant order, before averaging).

Two ways to create a metric

There are two paths to a new metric, and both produce exactly the same result:

The metric builder wizard (recommended for most people). Click Build on my site on the Metrics page and A vs B opens your own website with a guided panel on top. To track a button, you click the button. To track a page, you confirm the page. No code and no URL rules to learn. The wizard creates click and pageview metrics. See Metric Builder Wizard.
The advanced form. Click New metric to fill in the definition yourself. This path covers every metric type, including custom events and org-wide metrics, and gives you full control over matching rules. See Click Metrics and Pageview Metrics.

In this section

What Are Metrics?Understand primary vs secondary metrics (tracked for context, not the winner) and the metric types available.

Tracking EventsDefine click, pageview, and custom events once, then reuse them across every metric.

Metric Builder WizardCreate click and pageview metrics by pointing and clicking on your own site. No code needed.

Click MetricsTrack clicks on any element using a CSS selector.

Pageview MetricsTrack visits to a URL using simple, exact, substring, or regex matching.

Custom EventsFire metric conversions from your own JavaScript code using avsb.track.event().

Revenue on Custom EventsPass a revenue value with custom events to see Revenue Impact in your results.

Tracking RevenueRecord purchase events with avsb.track.purchase, auto-collect from GTM, or send from your server.

Revenue Deduplication and RefundsHow A vs B avoids counting an order twice, handles refunds, and works out net vs gross revenue.

Ratio MetricsTrack per-visitor ratios, like revenue per visitor, with a range you can trust.

Quantile MetricsMeasure the slowest or highest values, like page-load time, not just the average.

Composite MetricsCombine several metrics into one weighted score for experiments with more than one goal.

WinsorizationCap extreme values at a chosen percentile so outliers don't swing the test outcome.

Managing MetricsCreate, edit, delete, and reuse metrics across experiments.

Cross-Project MetricsShare one metric across projects and read its data from a single source project.

Key concepts

Every experiment must have exactly one primary metric (the one metric it's actually judged on). You can also add as many secondary metrics as you like. They give you extra context, but don't affect the winner calculation.

A metric belongs to your organization, not to one experiment. You create it once and attach it to as many experiments as you need. Each metric lives in a single project, or is shared org-wide across every project. Only an organization admin can create an org-wide metric. Any change you make to a metric applies to every experiment that uses it. See What Are Metrics? to learn more about where a metric can be used.