Does adding a trust signal help retested homepage form changes?

Our retested homepage form changes split by device, so we tried them again with the original mobile service picker, alone and with a trust signal. Over 15 days, the trust version was down 9.6%, short of our confidence bar.

Not significant Responsive 15 days Not called

SURFACE

Homepage

DEVICE

Responsive

VERTICAL

Ecommerce (B2C)

PRIMARY METRIC

Conversion rate

VARIATIONS

2

The hypothesis

By keeping the original mobile service picker and adding a trust signal to the retested form, we will keep the desktop gains and lift mobile, resulting in more completed orders.

In the previous round, desktop conversion rose and mobile fell, and the mobile drop pointed at the new service dropdown. We kept the original mobile selector this time. Trust messaging had also helped earlier on the homepage, so we tested adding a trust signal too.

What we changed

The retested form alone, or the same form with a trust signal

Var A repeated the previous round's changes, a new form headline and services moved below the fields on desktop, but kept the original service selector on mobile. Var B was Var A plus a trust signal near the form.

VAR A: BACK TEST

New form headline
Services moved below the fields on desktop
Original service selector kept on mobile

VAR B: BACK TEST + TRUST

Everything in Var A
Trust signal added near the form

The result

The trust version was down 9.6% over 15 days, short of our bar

CONVERSION RATE

-9.6%

VAR A CONVERSION

-23.2%

Over 15 days, Var B converted 9.6% lower than the original, and Var A converted 23.2% lower. Neither result reached our confidence bar, and each version had only a small number of conversions. Form submits were up 2.4% for Var B. Results swung during the run and then leveled out with no clear winner, so we paused the test.

What we would do next

Stop retesting old wins and test new ideas on this form

2 rounds of retesting earlier wins on the shorter form did not give a clear result, and traffic was too low to settle small differences. The trust signal version stayed closest to the original, so it is the better starting point if we test this form again.