Guest Checkout

Users had to be logged in to buy. At HSE's most critical moment in the funnel, that was a barrier. We removed it. +6.6% conversion uplift, A/B tested.

Overview

A focused checkout redesign that let users buy without an account. Multiple flow versions were tested qualitatively before shipping the A/B test that confirmed the uplift.

Guest checkout in action.

My direct contribution

Direction, coaching & critical decisions

  • Defined the strategic direction for the entry point framing: how to position "Continue as Guest" against "Log in" so the guest path never felt like a second-class option
  • Directed the key decision to move the account nudge from pre-purchase (where it killed conversion) to post-confirmation (where it performed); this reframing, developed with the checkout designer, was what made the final version work
  • Set the progressive data collection model: each step asks only what it needs at that moment, nothing speculative
  • Led the design critique sessions that killed two earlier versions before they went to qualitative testing

Team & collaboration

Mentorship, alignment & delivery

  • Coached and mentored the checkout team's designer throughout the project, guiding end-to-end flows, edge cases, and error states while keeping execution in the designer's hands
  • Worked closely with the checkout PM to ensure both business goals (conversion, revenue) and user goals (trust, clarity) were represented in every decision; also aligned Dev leads on payment integration constraints that shaped the design
  • Coordinated the qualitative testing sessions that determined which version went to A/B
  • Ran visual QA before release; checkout is the one journey that cannot break quietly

Duration: 2024 · Channels: Web and app · Measurement: Amplitude

Impact

  • +6.6% conversion rate uplift · A/B tested
  • +8.71% overall conversion improvement across checkout
  • +10.9% Add to Cart rate

The Problem

To buy at HSE, users had to be logged in. New customers had to create an account before they could proceed. Users who didn't remember their credentials had no easy path forward. The barrier was not protecting the business; it was costing it conversions at the worst possible moment.

The funnel data made the problem clear. Drop-off at the login/account step was measurably higher than industry benchmarks for comparable e-commerce journeys. We knew the "why" without needing research to confirm it: forced registration is a known conversion killer. Baymard Institute benchmarks the figure at 19% of adults who will abandon an order rather than create an account. What we didn't know was how to remove it without undermining the logged-in experience that existing customers relied on.

Three versions, two rejected

The team explored three approaches before running qualitative sessions. Two were killed before they reached users.

  • Version A, "Minimal prompt": showed a single "Continue" button with a small text link to log in. Users in critique sessions flagged that it was confusing: it wasn't clear whether they were buying as a guest or creating an account. Rejected.
  • Version B, "Explicit choice screen": full-screen split between Guest and Log In, with equal visual weight. This solved the clarity problem but made the guest path feel like a downgrade. In qualitative testing, users hesitated because they didn't trust whether their order would be "saved" or "real" without an account. Rejected.
  • Version C, "Progressive entry" (shipped): a clear primary path with "Continue as Guest" as the default action, and "Already have an account? Log in" as a secondary link. No forced choice screen. Account creation offered after order confirmation, not before. This version passed qualitative testing on trust, comprehension, and intent, and went to A/B.

Why the entry point framing was the hardest decision

The tension in this project wasn't technical; it was psychological. Labelling something "guest" implies impermanence. Early in testing, users interpreted it as "my order won't be tracked" or "I'll have no record of this purchase." That interpretation killed purchase intent even when the flow itself was simple. Checkout research describes this as perceived friction: uncertainty about what a label means creates the same cognitive cost as a genuine extra step in the flow.

The solution was to move the account nudge entirely out of the pre-purchase journey and into the post-confirmation state. Users who had already bought were far more likely to create an account to "save their order details", and far less likely to feel like they were being coerced. That single reframing was the insight that made Version C work.

Takeaway

The best checkout gets out of the way. Removing one requirement at the highest-stakes moment in the funnel, and reframing when you ask for account creation, produced a measurable, lasting result. Checkout optimisation is not cosmetic design; it is revenue recovery. Every barrier removed converts intent that was already there.