Testing the customer experience of Amazon's Prime Day

Posted on July 17, 2015
1 min read

Share

Prime Day was Amazon's global shopping event on July 15, 2015, that offered more deals than Black Friday, exclusively for Prime members. We decided to run a study to see how users would respond to Prime Day, and also to see if Amazon's design made sense to them or not. Here are three things we observed:

Observation #1: Lack of relevancy

Users were irritated by the random mashup of deals into big lists, and their inability to filter by categories that were personally relevant to them. Members—who theoretically use the site more often and cultivate wishlists, etc.—tended to be more irritated than non-members.

Observation #2: Unclear page elements

The problem above was amplified by the fact that once they left the big Prime Day lists, it was difficult to tell what did and didn't qualify for Prime Day. Users relied on the wrong page elements to determine if something was part of Prime Day. They either used:

  • The "Prime" icon, which just indicated if the product qualified for Prime 2-day shipping, or
  • The discounted price information in red, which, if it was a steep discount, indicated to users that it must be part of Prime Day

Observation #3: Social proof in action

Although users didn't respond particularly well to the Prime Day clock, the "% claimed" indicators, and the "Deal Over" information, many of them admitted that the "% claimed" indicator could influence their purchasing decision. Many users reported that they'd be less likely to purchase something with low % claimed (because it meant it was a bad deal), or they'd be more likely to purchase something with a high % claimed (because the deal must be incredible).

 

In this Article

    Read more

    • At Crafted Seattle, leaders explored why human understanding—not AI alone—will determine who builds the products people actually want.

      Blog

      The most valuable skill in the age of AI isn't what you think

      The race to build faster may be obscuring a more important question: Are we...
    • Team of UX designers optimizing websites and apps

      Blog

      How designers can validate AI-generated concepts faster with UserTesting’s MCP Server

      Nothing kills a flow state faster than having to constantly switch between different windows...
    • MCP with UserTesting

      Blog

      With UserTesting’s MCP server, researchers can stop switching between AI analysis and research tools

      You shouldn’t have to rebuild the same research workflow across five different tools just...