Tuft & Needle / 2018

Minimizing decision fatigue on product pages

One of the early sketches for a better product page

Hundreds of poorly differentiated products on the Serta and Beautyrest sites were creating a bad experience for customers.


We convinced Serta and Beautyrest leadership to simplify their product mix and group the remaining products using a configurator.

My Role

Strategy, UX Design, Prototyping, User Research

Tuft & Needle changed the mattress industry by creating a new way to buy mattresses. Our try-at-home, bed-in-a-box model with free shipping and easy returns raised the bar on customer support and convenience. And while we created an industry that has been steadily stealing away market share from the traditional players, there’s still a majority of shoppers who buy mattresses through conventional means. Our merger with Serta Simmons Bedding (SSB) is our attempt to change the way traditional mattress companies do business, making it a better experience for customers.

In preparation for the merger, we kicked off a preliminary study to determine the most effective strategy for making quick, positive changes to Serta’s and Beautyrest’s customer experience.

Identifying the problems

After conducting some initial research and an audit of Serta’s and Beautyrest’s websites, products, and services, we determined that several things hold the brands back from having the type of customer experience that modern shoppers have come to expect.

  1. Too many product offerings with unclear differentiation
  2. Cluttered, complicated websites with confusing information architecture and convoluted purchase paths.
  3. Poor customer support, especially around warranty claims and returns

The configurator: a simplification of choice

We had many multidisciplinary brainstorming sessions generating and evaluating various solutions. To address items 1 and 2, we decided that the best first step was to streamline the product offerings by removing hundreds of products from the online shopping experience and merging the best-selling SKUs into a few distinct, configurable products.

Taking inspiration from the automotive and tech industries, we experimented with creating product detail pages that would allow us to offer multiple variations of a product but display them to the user in a simplified, benefit-based decision flow. We called this new type of product page a “configurator.”

A wireframe for the first step of the product page selector
An initial wireframe. In this prototype, the screen’s right side would be sticky to the viewport as the left side scrolls.

One of the critical decisions was how to display the upgrades. Should users be forced to make a selection for each option? Or should the upgrades be checkboxes that could be easily skipped? There were pros and cons to both solutions.
A concept focusing on upgrade features and benefits. The “best for” section was intended to help reduce decision fatigue.
Another wireframe option that focuses on the benefits of the upgrades, not the features

Merging four models into one: the initial prototype

I conducted eight in-person tests with the initial prototype. In this version, I took four different products and packaged them as one product with modifications. In general, the feedback was that the proposed solution was intuitive and easy to use. Although each participant was technically choosing between 4 different models, the build-your-own mattress flow helped tremendously with decision fatigue. They didn’t feel that they were choosing between distinct models. They thought they were configuring one model with the features that made the most sense to their sleep preferences.

A screenshot of the initial wireframe prototype
A wireframe from the initial prototype. Because the questions we wanted answered were about flow and functionality, I created the prototype using wireframes rather than fully designed comps. Using InVision’s fixed component feature kept the right side sticky in the viewport as participants scrolled.

A second prototype, this time for Tuft & Needle

After exploring Serta’s configurator, leadership wanted to look at how it might function for Tuft & Needle. Using our existing product roadmap, I mocked up what this structure might look like for Tuft & Needle’s mattress lineup.

A reverse lander. After researching analogous industries for inspiration, the idea of a reverse lander was born. This benefit-driven narrative page would allow the user to get an overview of the mattresses’ various features and then dive in further to the sections that interested them.
Another wireframe option that focuses on the benefits of the upgrades, not the features
A screenshot of the initial wireframe prototype
The purchase flow, step 1. Once a user decides to purchase, the buy button at the top of the reverse lander would take them to the purchase flow. This flow would walk them through one choice at a time and update the mini-cart, sticky to the footer.
Step 2. For the T&N prototype, we wanted to test being more explicit that these were different models.
Another wireframe option that focuses on the benefits of the upgrades, not the features
A screenshot of the initial wireframe prototype
Step 3. The addition of a removable, washable cover.

Convincing SSB to take the leap

I created one other prototype for Serta that iterated on the previous designs and skinned it with their branding to help them better visualize it. We then presented the work to SSB leadership. Interestingly, seeing how the configurator and reverse lander might work for Tuft & Needle, a more successful ecommerce site, helped convince SSB of its potential for success on their sites. With these prototypes and research findings, we persuaded the SSB leadership team to move forward with this idea of configurable products.

This project laid the groundwork for the successful relaunch of custom ecommerce sites for both Beautyrest and Serta and is on track to change how both brands handle product strategy moving forward.

Thank Yous

JT Marino Strategy

Aaron Whitney Strategy

Tyler Marino Project Management

More work