Skip to content
Run an A/B test

You pick the idea.
Spar runs the test.

An A/B test used to need a developer, a statistician, and a guess about when to stop. Spar does all three. You pick the idea and drag a card.

Frontend variants ship through Spar. Backend changes still go to your engineers.

Variant code, written for you

Spar generates production-safe variant code from your idea. No engineer in the loop.

Live in a few clicks

Drag a card from Ready to Running and Spar deploys it to your store.

A winner at 95% confidence

Sequential significance declares the winner for you. No test duration to guess.

The test, running itself

Watch a card cross the board.

Every move between columns is work that used to need a person. Spar writes the variant, deploys it, runs the statistics, and flags the winner. You pick the idea and drag.

spar.so/your-store/board
Draft2
You set it up

Free-shipping bar on cart

Variation B

Target page and targeting confirmed.

Generating1
Spar writes the code

Trust badges under price

Variation B
Spar is writing the variant code
Ready1
Spar deploys variations

Bundle upsell on PDP

Variation B
Code ready to deploy
You drag
Running3
Spar runs and measures

Sticky add-to-cart on PDP

Variation B
95%
Day 618,420 sessions
You drag
Completed5
Spar calls it at 95%

PDP hero: benefit-led copy

Variation B won
Winner declared at 95%
Spar does it automaticallyYou pick and drag

Illustrative. Card names and numbers are sample values, with no real store behind them. The board, the columns, and the automation between them are real.

The seven steps

Spar does most of the work.

Here is the real flow, end to end. You pick, approve, and drag. Spar generates, deploys, measures, and declares the winner.

  1. Promote the opportunity

    You

    Open an active opportunity, review the test ideas Spar generated against your store data, and click Test on the one you like. Spar promotes it to the Tests page in Draft with the details pre-filled.

  2. Configure setup

    You

    Open the test detail to refine the idea description, confirm the target page and URL targeting, and add a variation. You set the direction; Spar takes it from here.

  3. Wait for the variant to generate

    Spar

    The status moves to Generating while Spar produces production-safe variant code. It usually takes a few minutes, then flips back to Draft, ready for your checkpoint.

    sticky-atc.variant.jsgenerated

  4. Move to Ready, then deploy

    YouSpar

    Drag the card to Ready for a final look, then to Running. Spar deploys the variant through whatever execution target you picked. No PR, no release queue.

  5. Monitor on the Analytics tab

    Spar

    Watch sample size, per-variant conversion rate, and a sequential-significance chart. Spar uses sequential statistical testing, so you never pick a duration up front. It flags a winner at 95% confidence.

  6. Declare the winner

    SparYou

    Spar flags the result when it lands. Drag the card to Completed to lock it in. From there, ship the winning variant as a Fix, archive if inconclusive, or re-run with a tweak.

  7. Capture the learning

    You

    Wins and losses both matter. Add a short note to the test, or pull it into a Notebook, describing what you learned for the next time a similar idea comes up.

Connected workflows

This is one step in the loop.

Start with an audit

Find the opportunities worth testing before you pick one.

Deploy through your own stack

Run the variant through GrowthBook, Intelligems, Convert, or the Shopify theme app.

Wrap results into a report

Turn a declared winner into a shareable readout for stakeholders.

Your next test is a card away.

Pick an idea Spar surfaced from your store, approve the variant, and watch it run. We handle the code, the deploy, and the call.

See an example site