A/B Test Strategy for Owala E-Commerce

UX Design

Spring 2026

UX Researcher / Strategist / Designer

Photoshop (mockups), Miro (peer feedback), WCAG contrast checkers (Chrome Plugins), Excel (audit data), Google Slides (Presentation)

See Presentation

Audit Owala's e-commerce product pages and propose two A/B tests — not just design opinions, but tests with a measurable hypothesis, a KPI, and a rationale a stakeholder would accept.

Before touching the product page, I did the work most designers skip: a brand assessment covering competitors, SWOT, target audience, and the full customer journey. That's where I caught the first issue — buried in a routine content inventory, not a flashy discovery.

Issue 1 — Trust signals were invisible. Product feature callouts (the icons customers use to quickly validate "does this solve my problem?") were positioned far below where a shopper would actually need them to make a decision.

Issue 2 — Inconsistent, low-contrast product backgrounds. Some pages used neutral gray behind the product; others used a tint of the product's own color, which often failed WCAG contrast minimums. I ran contrast checks myself: most pages failed AA, none met AAA.

Test 1 — Trust Signal Placement

Hypothesis: Moving trust signals higher on the page reduces purchase hesitation.

KPIs: Purchase rate, cart abandonment, returns.

Method: 50% traffic split between current placement and elevated placement.

Test 2 — Background Consistency & Accessibility

Hypothesis: A uniform neutral-gray background improves color accuracy perception and meets AA contrast — reducing color-related returns.

Reasoning: A gray backdrop lets the product's true color read accurately, rather than being visually distorted by a tinted background — directly addressing why customers return items over color mismatch.

KPIs: Bounce rate, purchase rate, returns attributed to color.

This wasn't just "make it prettier" — it connected an accessibility failure to a business metric (returns), which is the exact translation skill non-design stakeholders need to see to trust a designer's recommendation. I also validated the trust-signal comp through peer critique in Miro before finalizing the test design, incorporating feedback like simplifying visual clutter around the product description.

A stakeholder-ready presentation proposing two tests, each with a clear KPI, sample size, and expected business impact — moving accessibility from "nice to have" to "measurable revenue lever."