Experiment Lifecycle

Every experiment in Split Test Pro moves through a small set of states. Knowing what each state does — especially what it does to traffic and data collection — saves you from accidentally invalidating an experiment with a wrong-button click.

The Five States

State	What it means	Variants shown?	Data collected?
Created	Draft. You’re still configuring it.	No	No
Started	Live. Visitors are being assigned and tracked.	Yes	Yes
Paused	Temporarily stopped. Existing assignments stay sticky.	No (new visitors see control)	No (new data)
Completed	Manually ended. Results are frozen as a final snapshot.	No	No
Archived	Soft-deleted from the main list for organization.	No	No

State Transitions

Created → Started

Clicking Start Experiment runs the pre-launch checklist and, if you confirm, transitions the experiment to started. From this moment, the variant injection script picks it up on the next pageview and begins assigning visitors.

Started → Paused

Click Pause to halt new traffic. Visitors who were already assigned a variant don’t lose their assignment — the cookie persists — but no new visitors are bucketed and no new events are recorded against the experiment. Pause is reversible; you can resume by clicking Start again.

Started → Completed

Click Declare Winner (available once a variant crosses the 95% probability-to-be-best threshold) or end the experiment manually. The experiment moves to completed. Results are frozen — the dashboard shows a “Final snapshot — results frozen” banner — and no new data is collected. See Declaring a Winner.

Any state → Archived

Click Archive to hide an experiment from the main list. Archived experiments are still accessible (toggle the filter on the Experiments view) and their results are preserved. Use archive to keep your active list clean once you’ve extracted the learnings.

Completed vs Archived

These two states sometimes get confused, but they’re different tools:

• Completed is about the experiment itself — it ran, it ended, results are final. You can compare a completed experiment’s result to others when planning the next test.
• Archived is about your workspace’s organization — hide tests you don’t need to see in the active list anymore. Most teams archive experiments only after they’re completed, but you can archive a draft you decided not to ship, too.

A typical workflow: experiment runs → completes → sits in the “Concluded” section for a sprint or two while the team applies the learnings → gets archived once it’s no longer top-of-mind.

Copying an Experiment

You can duplicate any experiment — running, completed, or archived — to use as the starting point for a new test. From the experiment detail view, click the more-actions menu and select Copy. The duplicate is created in created state with all variants, targeting rules, and metric configuration cloned. Adjust as needed and start it fresh.

This is the right pattern for iterating on a winning hypothesis: don’t restart a completed experiment, copy it and tweak.

Common Mistakes

• Pausing instead of completing a finished test. Pause leaves the experiment in your active list and looks “incomplete.” If you’ve decided on a winner, click Declare Winner (or end manually) — that’s the right state.
• Re-starting a paused experiment with a changed variant. That’s effectively a new experiment with a different population. Copy it instead.
• Archiving without completing. A running experiment that gets archived is still consuming traffic and recording data — archive doesn’t pause it. End the experiment first, then archive.

Next Steps

• Manage how visitors are split between variants: Traffic Allocation.
• Run several tests at the same time without stepping on each other: Running Multiple Experiments.
• Learn the rules for ending a test confidently: Declaring a Winner.