# A Sticky Buy Button on Long Product Pages: A/B Test

Page: https://alidemirbas.com.tr/en/lab/ab-test-library/tests/a-sticky-buy-button-on-long-product-pages-ab-test

## Hypothesis

If the purchase CTA stays fixed as visitors scroll a long product page, they can buy from wherever they are while reading the description or reviews. Expect a higher purchase conversion rate (CR) per assigned visitor if that holds.

## When to run this test

On long product pages, heatmaps with heavy attention low on the page but few trips back to the button, or session replays of visitors who read the reviews and leave, suggest people give up as they scroll away from the buy button. If the idea is right, those visitors move to purchase without scrolling back up, and completed orders rise. If the page is short or visitors already return to the button, access is not the problem, so review price, stock status and product information first.

## Primary metric

**Purchase conversion rate (CR)**: Compare completed orders per assigned visitor with and without the fixed purchase CTA.

## Other metrics

- **CTA click-through rate**: Use clicks as a diagnostic measure.
- **Add-to-cart rate**: Check whether shoppers take the first purchase step.
- **Page exit rate**: Compare exits in the same session window.

## What to check before and during the test

- **Sticky behavior**: Keep the existing purchase control visible using one predefined sticky rule.
- **Action**: In both arms, the button must lead to the same point in the order flow and preserve product selection and validation.
- **Visitor history**: Review new and returning visitors separately.
- **Appearance**: Keep label, size, color and price information unchanged.
- **Trigger**: Set when the sticky behavior begins before launch.

## What to watch when reading the result

- CTA click-through rate only shows the raw click gap; clicks that do not turn into orders give you nothing to ship on. Base the call on purchase conversion, use add-to-cart rate to see where shoppers stop, and watch page exit rate, since a rise may mean the fixed button is getting in the way.
- This sticky rule, its trigger point and long product pages are what the result covers. Read new and returning visitors separately, because people who know the page may already reach the button easily, and validate again before using it on shorter pages.
- During QA and again while the test runs, check that the button does not cover much of the screen, does not sit over important content and does not block the view when the mobile keyboard opens. Keep shadows and animation restrained, or the difference may come from an eye-catching effect rather than the fixed position.
- A small or unclear difference does not show that the sticky button does nothing, only that this test could not tell the arms apart. Test the gap within new and returning visitors against chance with the [A/B test significance calculator](https://alidemirbas.com.tr/en/calculators/ab-test), one group at a time.

## Frequently Asked Questions

### Does the sticky button lead to the same step in both groups?

Yes. Preserve its destination, action, product selection and validation. Only the existing purchase control staying visible changes. Use purchases per assigned visitor as the primary outcome.

### How can I detect disruption from the sticky button?

Monitor exits. Avoid an oversized button or overlap with important content, and check the layout when the mobile keyboard opens.

### Can I use one result for all visitors?

Examine new and returning visitors separately. Effects may differ. Keep the point where sticky behavior starts unchanged throughout the test.