Secondary Metrics
Secondary metrics are additional measurements that run alongside your primary metric. They do not determine who wins — only the primary metric does that — but they provide the context you need to understand whether a winning result is genuinely good for your business.
How they differ from the primary metric
Your primary metricis the one measurement that determines the winner of your experiment. When a variation's winning probability on the primary metric crosses the configured threshold, that variation is declared the winner.
Secondary metricsare measured and displayed alongside the primary metric, but they have no bearing on the winner calculation. Think of them as a health check on your primary result. They help you answer the question: "Yes, we won on the primary metric — but did anything else break?"
Viewing secondary metric results
On the Results page, secondary metrics appear in collapsible sections below the primary metric section. Each section shows the full variation performance table — visitor counts, conversions, conversion rate, improvement vs control, credible interval, and probability to beat control — for that specific metric.
Expand a secondary metric section by clicking on it. The data shown is the same format as the primary metric, making it easy to compare performance across metrics.
Use cases and examples
Funnel metrics
A common setup is to use a bottom-of-funnel metric as the primary (such as "Purchase completed") and add upper-funnel metrics as secondary (such as "Add to cart click" and "Checkout page view"). This lets you see whether a winning result on purchases was driven by improvements across the full funnel or by a change at just one step.
Example interpretation: if purchases improved 15% and add-to-cart clicks improved 18%, the change likely improved intent to purchase across the board. If purchases improved 15% but add-to-cart clicks were flat, something about the checkout flow — not the product page — drove the improvement.
Guardrail metrics
Secondary metrics are excellent guardrails. Use them to monitor metrics that should not get worse, even if you are not trying to improve them.
Example: your primary metric is "Sign-up form submission." You add secondary metrics for "Bounce rate (pageview)" and "Time on page." If your primary metric wins but the bounce rate goes up significantly, that is a signal that the variation may be pushing visitors to convert too aggressively — at the cost of the experience for visitors who are not ready to sign up.
Revenue alongside conversions
If your primary metric is a conversion event without revenue data (such as a click or a pageview), you can add a revenue-enabled custom event as a secondary metric to get an approximate revenue view. For example: primary = "Checkout page visit," secondary = "Purchase completed with revenue." This gives you both a fast-moving primary metric and a revenue impact view.
Adding secondary metrics
Secondary metrics are added in Step 4 (Metrics) of the experiment builder. After selecting your primary metric, click Add secondary metricand choose from your project's metric list. You can add as many secondary metrics as you like, though keeping the list focused (3–5 metrics) makes results easier to interpret.