Know what wins before you commit to it

Server-driven split tests on anything you already flag. Define variations, weight the traffic, pick the goals, and watch the winner emerge in real time.

Split Testing in SessionSight runs server-driven variation tests with traffic weighting, goal tracking, and automatic winner detection, so you know what wins before you commit to it.

Every test you've run, still running, or waiting to ship

Split tests live next to feature flags and reuse the same value types: ID (route to a different component), TEXT (swap a headline), or JSON (reconfigure a whole section). Everything server-driven, so there's no client-side flicker.

Split Testing

Run server-driven variation tests to optimize conversions

+ New Split Test
Total Tests
4
Running
2
Completed
1
Draft
1
All Draft Running Paused Completed
Name Status Type Variations Traffic Started
CTA Button Color cta-button-color
DRAFT ID
Green (Current)Orange
50% -
Hero Headline Test hero-headline
RUNNING TEXT
ControlVariant A
100% Apr 6, 26
Checkout Flow Experiment checkout-flow
RUNNING ID
Control (Multi-step)Single PageAccordion Steps
80% Mar 30, 26
Pricing Page Layout pricing-layout
COMPLETED JSON
ControlTwo Column
100% Mar 6, 26

Define variations once, let the server handle routing

Each variation gets a key, a human-readable name, a weight for traffic distribution, and a value that matches the test type. Boolean, number, text, JSON, or an ID for component-level experiments. All values land in the same SDK call as your feature flags.

Variations

Define the variations visitors will see.

+ Add Variation
CONTROL Control
Key
control
Name
Control
Weight
50⌃ ⌄
Value
See your users, stop imagining them.
Variant A
Key
variant-a
Name
Variant A
Weight
50⌃ ⌄
Value
Conversion tools teams actually ship with.

Pick the outcomes you're optimizing for

Goals pulled from your goal library

Select any goal you've defined and it becomes a conversion metric for the test. One gets marked Primary; the rest become secondary metrics, measured in parallel so you know whether a win on signup came at the cost of revenue.

Goals

Select goals to track conversions for this test.

Newsletter signup Primary
Starter plan purchase Set primary
Feedback submitted
Daily Signups
Weekly Purchases
Monthly Revenue Set primary
Quarterly Active Users
Yearly Total Orders
Monthly Page Views
Weekly Support Tickets Resolved
Monthly Trial Conversions

Ramp it up, ramp it down

Traffic Allocation controls what percentage of eligible visitors join the test. Start at 20% for a safe read, move to 100% when the numbers hold. Change it live without redeploying.

Traffic Allocation

Percentage of eligible visitors included in this test.

57%
10% 25% 50% 75% 100%
Eligible visitors 12,482 / day
In this test 7,115 / day
Projected to significance 8 days

Stop debating the headline,
ship them both

Server-driven split tests let you run variations in production without redeploys or client-side flicker.

Start your first test
Keep exploring

Three more tools worth a look.