Every test type, one engine.
A/B, split URL, multi-page, multivariate and auto-allocated tests on any site; price, offer, shipping and template tests on Shopify. Every type shares one visitor assignment and one set of statistics.
On any site
Shopify, or any store with the Benchmyrk script tag.
- A/B and A/B/n testsChange copy, images, layout or styles on the live page and compare versions.
- Split URL testsSend part of your traffic to a different page and compare whole pages.
- Multi-page testsOne change across product pages, cart and collections, one version per shopper.
- Multivariate testsSeveral elements at once, in every combination, with interactions checked.
- Auto-allocation (bandits)Move traffic toward the better variant while a campaign runs.
On Shopify
These change what checkout charges or which template Shopify renders, so they need Shopify.
- Price testsShopifyEach group sees and pays its own price, charged by a Shopify discount at checkout.
- Offer and discount testsShopifyPercent or amount off, spend tiers or buy X get Y, given at checkout.
- Shipping testsShopifyFree-shipping thresholds and shipping discounts on your own rates.
- Template testsShopifyAlternate templates of your live theme, with no theme edits.
Questions
Which test should I start with?
Start where traffic and revenue are: usually an A/B test on your best-selling product pages. Price, offer and shipping tests change what shoppers pay, so plan them with your margins in mind.
Can I run several tests at once?
Yes. Tests on different pages run side by side, and tests that would reach the same visitors can be kept apart with mutually exclusive groups.
Do Shopify-only tests work on every Shopify plan?
Yes. Price, offer and shipping tests run on a Shopify discount function, which public apps can use on every Shopify plan, and template tests use Shopify's alternate templates.
Not sure which test to run first?
We'll look at your store and traffic with you and suggest where a test is most likely to pay off. Or install the Shopify app and start with an A/B test.