CRO Experiments

A/B testing aim to improve conversion rate

Part of Conversion Rate Optimization (CRO) work is A/B testing experiments by making small adjustments throughout the platform grounded in thorough data, research and industry best practices to see how customer responses and their impact to help increase their conversion through the funnel.

Requirement

requirements are given by the CRO specialists, which outline the requirements and features to be redesigned. These requirements serve as the foundation for thorough design research. This research includes past experiments analysis and UX guidelines like Baymard.

These recommendations undergo multiple discussions with stakeholders, such as the BU and developers, to evaluate the feasibility and assessment of the impact A/B test would have on the current website to prepare for testcase when doing QA.

Design

We try to find a compromise between best practices, BU requirements, along with brand CI guidelines while still making small but impactful changes. Since Central Online sells a variety of products, different scenarios and states have to be considered.

Example 1: image overlay

Overlays are not recommended as they obstruct the product image. However, they are useful for BU to communicate promotions.

Design suggestion: reducing the size of the overlay while maintaining the same color and font to ensure familiarity.

Example 2: brand navigation

Brand logo will improve product navigation and brand visibility. However, due to BU limitations and the extensive brands Central has, brand Logo cannot be provided.

Design suggestion: use brand name instead for quicklinks instead of logo

Design QA

Once the design has received approval and gone through development, Design QA will recheck for design accuracy, user experience and potential impact the customer may face with different use cases. During this phase, we carefully check the design to ensure that all aspects, including color, typography, and spacing, align are all precisely with our original design specifications for both Thai and English.

Example 3: Product description length

Excessively lengthy product description impacts the customers' scroll depth as observed from the scroll map. To address this, we introduced a 'see more' option to condense the product description while keeping certain portions visible.

Design analysis

At the conclusion of each cycle, each experiment is successful, a failure, or inconclusive. Designer and analyst will do an analysis of each design, using heat maps and click data, to see how customers interact. these analysis then helps us better create designs for customers backed by data to improve conversion rates.