Isriel Robinson  ·  CRO & Experimentation

Step One Life  ·  US  ·  Oct 2025  ·  3-variant ladder

How much trust signal
is enough

The PDP lead image carried only product; the trust assets sat below the fold. A three-variant ladder — Trustpilot, +render, +USP icons — asked how much signal pays, not just whether it works. +8.5% conversion, implemented, with the AU replication queued.

+8.5%

conversion rate, winning variant

3

variants laddered by information density

15

days, US storefront

AU

replication queued on the trusted stack

Role
CRO Lead + Ecom Ops — ladder test design, cross-market replication plan
Programme scale
Control + 3 variants, US PDP lead image
Primary metric owned
Conversion rate
Status
Implemented in the US; AU rerun queued (US result from a CRO tool under trial — treated as directional until replicated)

Contents

  1. 1. The Problem
  2. 2. The Test
  3. 3. The Result
  4. 4. The Caveats, Stated Plainly
  5. 5. Key Insights

01The Problem

The PDP lead image is the most-viewed piece of real estate on the site, and it was carrying only product. Meanwhile the trust assets that answer a first-time buyer's actual question — is this a real company with real reviews? — sat further down the page where a share of visitors never reach.

Hypothesis: if trust symbols and USPs are added to the lead image, conversion will increase because value is perceived immediately, before any scrolling.

02The Test

Rather than one variant against control, the test ran as a ladder — three variants of increasing information density on the lead image:

  1. Trustpilot rating alone
  2. Trustpilot rating + product render
  3. Trustpilot rating + render + USP icons

That structure answers a better question than "does trust work?" — it asks how much signal is enough, and where added density stops paying.

03The Result

+8.5% conversion rate for the winning variant, US storefront, 10–25 Oct 2025. The variant was implemented, and the concept directly seeded the follow-up backlog: a Trustpilot overlay on the first gallery image queued for the AU storefront.

04The Caveats, Stated Plainly

Two reasons this result is presented with a leash. First, the test ran while trialling a new CRO tool on the US site — results from a platform under evaluation get treated as directional until replicated on the trusted stack. Second, a US result is not an AU result; the plan documented at the time was to re-run in AU once the whole-theme tests settled.

That is how single-market wins are handled in this programme: implemented where proven, queued for replication where not.

05Key Insights

Trust belongs at the point of first judgement, not at the bottom of the page. The lead image is where the credibility question is asked, so that is where the answer has to be.

Test the dose, not just the drug. Laddering Trustpilot → +render → +USP icons turned a yes/no test into a design specification: it identified the density that pays, instead of landing on "more is better" or "less is more" by taste.


Source: experiment US_PDP_CVR_251010, Step One US storefront, 10–25 Oct 2025. Run during a CRO-tool trial; AU replication queued in the test backlog. All figures traceable to the named experiment ID.