Understanding Experiments
The Experiments page (labeled "Experiments" in the sidebar) is where you create, manage, and analyze A/B split tests. It provides a centralized view of all your experiments with key performance metrics.
Accessing Experiments
Log in to your Humblytics account and select your site.
Click Experiments in the sidebar under Analyze.
Summary Cards
At the top of the page, four cards give you a quick overview:
Active Tests — Number of experiments currently running
Total Visitors — Combined visitor count across all experiments
Avg. Lift — Average conversion lift across completed experiments (shown as a percentage, green for positive, red for negative)
Success Rate — Percentage of completed experiments that achieved a statistically significant positive result
Status Filters
Filter your experiments by status using the pill buttons:
All — Shows every experiment regardless of status
Drafts — Experiments in setup that haven't been launched yet
Active — Currently running experiments collecting data
Paused — Experiments that have been temporarily stopped
Completed — Finished experiments with results
Archived — Old experiments moved to archive
A search bar lets you find experiments by name.
Experiment Cards
Each experiment is displayed as a card showing:
Name — The experiment name
Status badge — COMPLETED, ACTIVE, PAUSED, or DRAFT
Creation date — When the experiment was created
Goal — The conversion goal (e.g., form submissions, click-through rate, revenue, bounce rate, session time, page reach)
Variants — Control and variant page URLs shown as pills
Visitors — Total visitors assigned to the experiment
Lift — Percentage change in conversion rate (green for improvement, red for decline)
Confidence — Statistical confidence level of the result
Creating a New Experiment
Click the New Experiment button to start the experiment creation wizard. For detailed setup instructions, see the Split Testing Overview.
Click Plan a Test to use the A/B Test Planner tool, which helps you calculate sample size requirements and estimated test duration before launching.
For more details on split testing, see:
Last updated