Skip to content

Test archive / Homepage

Does zip code autofill lift conversion on a homepage form?

The desktop homepage opened with a short form that asked visitors to type in their zip code. We tested filling it in for them based on their location. Conversion rose 8.9% over 9 days, but the test was paused before it could be called.

Not significant Desktop 9 days Not called

SURFACE
Homepage
DEVICE
Desktop
VERTICAL
Ecommerce (B2C)
PRIMARY METRIC
Conversion rate
VARIATIONS
1

The hypothesis

By filling in the zip code from the visitor's location, we will cut the effort needed to start the form, resulting in more visitors reaching the confirmation step.

The homepage form is the first step of the order, so every field is a small hurdle before the visitor sees any value. The zip code is the only field we can fill in for most visitors without asking, based on their location.

This test was part of a round focused on cutting effort in the first steps of the funnel. The same change ran as a separate mobile test at the same time.

What we changed

The zip code field arrived filled in from the visitor's location.

In the control, visitors typed their zip code into the homepage form themselves. In Var A, we looked up the zip code from the visitor's location and placed it in the field before they reached it. Nothing else on the page or in the form changed.

CONTROL
Control wireframe
VAR A: ZIP CODE AUTOFILL
Var A: Zip code autofill wireframe
  • Zip code filled in from the visitor's location
  • Rest of the form and page unchanged

The result

Conversion rose 8.9% in 9 days, short of a call.

CONVERSION RATE
+8.9%
REVENUE PER VISITOR
+0.4%

Var A lifted conversion 8.9% over control, and the lift held at each step of the funnel, so the filled-in field did not confuse anyone. Revenue per visitor barely moved, at +0.4%. We planned a full 14 days but paused at 9 days to start the next round of tests, and the result never reached our confidence bar.

The same change on mobile showed too small a lift to call. The change was not implemented.

What we would do next

Rerun it for the full 14 days before judging the lift.

The lift was steady through every step of the funnel, which points to a real drop in effort. But 9 days was too short to trust it, and flat revenue per visitor meant the extra orders were not adding value yet. We would rerun it for the full 14 days, alongside other fields we can fill in for the visitor.

Let us find where your conversion opportunities are.

Request an evaluation