Read an Experiment Report

Overview

Every experiment has its own reporting page, opened by clicking on the experiment from your dashboard. This article walks through each section.

Experiment Summary

The first section shows how the experiment was configured:

• Device Type - which devices are eligible.

• Traffic allocation - what percentage of total traffic is included.

• Segment - which visitor group is eligible.

• Created on - creation date.

• Reset on - the last time data was reset.

• Expires on - when the experiment ends.

• Paused on - the last time it was paused.

Below that, Traffic allocation for each Variation is shown, with a Preview button for each. Further down, the experiment's Audience (included and excluded pages), Exclusivity setting, Data collection exclusions (excluded IPs), and any Tags applied to the experiment.

Sample Size Calculator

Shows how much longer the experiment needs to run before reaching statistical relevance. See Use the Sample Size Calculator for the full explanation.

Performance

This is the core results section - goal data for the Control and each Variation, filterable by conversion.

The Main Goal (usually Sale) appears first. Each row shows the Control and Variations, along with their Experiment ID and Variation ID.

Understanding the columns

• Users - unique visitors who saw the experiment.

• Views - total times the variation/control was seen.

• Conversion rate - conversions ÷ users × 100.

• Conversions - number of visitors who reached the goal.

• Chance to win - statistical relevance; the probability the Variation genuinely outperforms Control. See Understand Statistical Significance & SRM.

• Total revenue - total revenue from visitors influenced by that Control/Variation who reached the Sale goal.

• Revenue/user - total revenue ÷ number of users.

• AOV (Average Order Value) - average revenue the experiment brought in per order.

Worked example

If the Sale goal shows a 1.57% conversion rate on Control and 2.64% on Variation, that means for every 1,000 users, 15 converted on Control and 26 converted on Variation.

Every Variation is measured relative to Control - an uplift is shown in green, a decrease in red, both as a percentage relative to Control.

Chance to win results are calculated using both Frequentist and Bayesian methods. A result above 95% on your main goal is generally considered a good signal to implement the change.

Press the Graph icon next to any goal to see its evolution over time, viewable as either Conversion Rate or Conversions.

Reporting for Surveys

Survey reports share the same Filter by controls as other experiments (Devices, date range).

The KPI cards shown are:

• Users - total visitors who saw the survey.

• Respondents - how many of them answered at least one question.

• Response Rate - respondents ÷ users, as a percentage.

• Average duration - average time to complete the survey.

• Leads - number of leads collected, if the Lead Collector is enabled.

• Lead rate - leads ÷ respondents, as a percentage.

Filtering answers

Use Filter answers to show only responses matching a specific answer on a unique-answer question, then press Apply filters.

Viewing answers

Each question's answers are shown as a bar chart, with the percentage and count of respondents who chose each option.

[Editor note: no matching screenshot exists in the source articles for this (Answers section bar charts per question). Needs a new screenshot.]

Filters

Devices

Filter results by device type, since experiments can behave differently across devices.

Attribution model

Choose between two models:

• Linear distribution - every touchpoint in a conversion path shares equal credit. If an A/B test and an Overlay both contributed to a $300 sale, each gets $150 attributed.

• Assisted conversion - each touchpoint gets full (100%) credit independently. The same $300 sale would be counted fully for the A/B test and fully for the Overlay - meaning the combined attributed total ($600) is higher than the actual sale, since both experiments are each treated as fully responsible.

Filter by goal

Build detailed filters based on specific goals and their values.

Display goals

Choose which goals appear in the Performance tab.

Timeframe

Choose which days are included in the displayed data.

Data refresh rate

Experiment data refreshes in real time (every second) while the experiment has fewer than 100 total users. Once it surpasses 100 users, data refreshes every 1-2 hours instead.

Exporting data

Download CSV

Three export options, found at the top right of the page:

• Download report (.csv) - the data as shown on the Performance page.

• Download report detailed conversions (.csv) - includes per-conversion detail, such as transaction IDs for the Sale goal.

• Download main goal conversions (.csv) - conversion data specifically for the experiment's main goal.

Reports

Generate a report from the experiment's current data, or download a previously generated one.

Notes

The final section shows any notes your team has added to the experiment, along with recent changes made to it - useful for tracking what was modified and by whom.