Online Wedding Rings

Concept Testing

What’s interesting

Rapid Iterative Concept Laboratory

With a small team of two designers and 1.5 researchers, we were able to generate rapid insights every week, creating a dynamic cadence of exploration, design, modeling, user insights, stakeholder feedback, and refinement. I strategized the cadence in a way that would be scalable on this and future projects to fit within a SCRUM sprint cycle, effectively providing a concept laboratory and insights engine.

Unmoderated Online Testing – Every week we used usertesting.com to launch several quick-turn unmoderated tests. Strategically, we chose to use this method for things such as baseline testing, competitive tests, and card sorting.

We were also able to explore sentiment around some parallel industry concepts. For instance, what about Warby Parker, Stitch Fix, or Tinder could be applied to wedding ring purchases?

In Person Testing – Once we ramped up to the 3rd week (buffering 2 weeks for recruitment), we began a weekly cycle of in-person testing. We were able to discuss big ideas and get fast insights from our prototypes and even some imperfect static models.

These tests were successful because of a close partnership between designers and researchers. Designers attended user interviews, and helped researchers spot additional opportunities.

After purchasing an engagement ring, only 10% returned for their accompanying wedding bands.

Concept: Your Team of Experts – Create a jeweler experience, putting clients at ease, helping them choose their style, and giving them confidence to purchase online.

Site styling was outdated, sales were lagging, and investors were pressing for an immediate refresh to stay competitive.

Calibrate screen: Users can measure a physical object if their device has first been calibrated. However, participants found it hard to understand why they must first measure a credit card or coin.

Pinch / Zoom or Dial: Participants reacted with ease and confidence to a simple size dial. Despite its ease, participants still felt they might need to visit a store to confirm ring size.

Printed Sizer: Participants had very low confidence in this method, citing issues with printers and accuracy of measurement as barriers to trust.

AR Measurement: Participants felt this would be desirable and effective if the simultaneous coin and ring finger measurement were accurate.

In-line help: With relevant content, users found these “tap to learn” boxes useful, answering questions before they even asked.

All-purpose “i”: Tapping “i” at any time would expand to show chat, call, and curated help topics. Users were delighted by this approach and found it discoverable.

A critical need for improving the 10% attach rate for engagement rings was to offer an intuitive ring matching assistant. The site had a cumbersome tool with flawed logic and numerous dead-ends and null sets. Our redesign proved intuitive and effective.

To move beyond filters to the service level of a personal jeweler, we tested a panel of intuitive wizards to coach style, fit and other aspects of the shopping experience.

Swipe Left: Users swiped left on the Tinder model, feeling it was perhaps not fitting or familiar for a site solidifying long term engagements.

A human coach makes customers for life: Every time we provided simple visuals indicating that a live representative could provide advice or coaching, our participants had highly positive reactions. Our client agreed they could prominently offer and staff personal advisors, and would explore the option of photograph and recommendation exchanges.

Findings: The Luxe Experience

Airbnb offered a familiar model for discovering and browsing targeted, curated lists in an intuitive way. Displaying curated content in sliders or carousels, with only 3 to 8 items each, allowed easy exploration and discovery.

Curated Carousels: Previously buried or unwieldy, we brought consistency and simplicity to help surface these lists.

Highly sensitive Home and Segment (Landing) pages were a critical part of modernizing the site. We prototyped in medium fidelity to present numerous interaction patterns, then skinned in high-fidelity for styling sentiment tests.

For the Catalog and Product Detail pages, two versions of each were designed in high fidelity and tested for appeal, discoverability, and readability, with focus on details such as priority of image & text, and patterns of browsing and selection. A final hybrid version of each was produced and launched.

Biggest Challenge

As lead researcher on this project, I collaborated closely with the rest of the design team to align each prototype with an effective test scenario, mapping screens & content to key tasks to be tested.

The first round of ideation / prototyping / testing was a bit bumpy, as our prototyper was scrambling late the night before in-person testing to finalize and debug their prototype. That meant that several of us had to then test to ensure that the final model still aligned with the morning’s script.

In all future 3 week (overlapping) design/test cycles, we added earlier checkpoints and script alignment and built in a weekly practice to shake-out the final script and prototype the day before we met with test subjects in person.

Notes in the cloud: Notes were compiled in a shared doc. The shell document was created before the meeting to align with key questions and predicted responses.

Designer huddle: Our note taker was a design team member, so we could do a quick huddle after each session to surface and align on the key observations.

Design impacts: In this lean research, we had to think weeks ahead. Today’s test’s design insights and client decisions may impact next week’s design.

Publish to the cloud: We could publish more quickly and at lower reporting fidelity than some other projects because our client and design team were both closely embedded in our testing process.

Visually organize by flow and feature: Having just tested from a familiar prototype, we arranged our results to follow a visual walkthrough. A table was created with each item given a visual snapshot or short video.

Visual walk-along: Often we had the prototype ready for walkthrough, so we could display findings in our table, then demonstrate on a connected mobile.

Positive / Negative / Recommendation: To give a full understanding, each option was examined for both positives and negatives. Ultimately the client needed a recommendation, which sometimes spanned into design and interaction guidelines, or simple “go” or “no” recommendation on capabilities.

Supporting Evidence: Qualitative evidence was critical to help participants make effective choices, especially when deeply held beliefs were challenged. Video clips and quotes were readily available, and qualitative evidence (ie 5 of 5 participants) was also surfaced. This was especially helpful around recommended changes to filter design and IA.

We created a prioritized list of dozens of potential scope items for future work, listing the recommended research questions, recommended approach, and desired outcomes of each proposed effort.

At present, the client is still implementing portions of the base study, but is actively considering a phase 2 project which will include some of these items.

Reflections

Rapid Concept Iteration: 

• Constant Insights – While this engagement was only for exploration of 6 concepts over 8 weeks, we proved that a small team could be setup within a company such as our client’s and constantly generate innovations and insights.

• Works within SCRUM – Once the pattern was established, we saw that this could be replicated within a SCRUM practice. High potential insights could proceed into a conceptual development and testing program, with further design refinement and partnership.

Concept Groups: 

• Families of concepts – The early research performed by a partner team provided deep insight into the the experience of various key personas, mapping highs and lows and identifying goals, needs and pain points. These formed clusters of opportunities, at an appropriately high level that left our team flexibility to discover new ways to meet those needs.

Ideation & Prototyping: 

• Close Calibration with Research – The project took flight once we reached our second design concept, after implementing lessons learned from round 1. Simply making mid-cycle alignment a priority, and targeting completion of the prototype a day early to allow a dry run and adjustment made all the difference.

• Prototype Fidelity – We didn’t always need the same fidelity. It was important to not over-design something that was intended to prove out only a high level interaction. However, there were certain experiences (such as the Luxe Experience) which demanded high fidelity, as “blue box” style wires created false signals. (“Should I comment about whether I like this blue? I don’t.”)

In Person vs Unmoderated

• In person testing empowers discussions – In person sessions were saved for nuanced discussions. We wanted to expose enough realism to trigger the imagination and a good conversation. But participants were more than willing to use their imagination with a statement such as “In real life this might …”

• Unmoderated for quick insights from existing sites – In our work we were able to steer our more creative efforts by keeping ideas grounded in user insights. Although usertesting.com was not perfect, it empowered us to obtain insights within hours on any question we could pose clearly. We chose to use it mostly on baselining and on comparative studies. Even more interesting, though, were answers to questions like “Go to warbyparker.com and explore their eyeglass shopping process. What if a similar site existed for buying wedding rings?”

Each Test Influenced the Next 

• Continuous learning – If we explored 3 to 5 options on a certain element one week, the winner of that exploration would show up as a background element in the next. So for instance, we explored different modes of help, aligning on two methods. We had an opportunity to include in-line tips and a floating “i” in the background as we explored filter options the following week. This provided a secondary signal to help triangulate success or failure.

Surfacing Results & Recommendations: 

• Regular frequent stakeholder reviews – Our stakeholders were eager for Friday afternoons. Exciting insights were delivered every week and they were hooked, wanting to find out the results. Keeping them this connected from start to finish on this project provided them a very real sense of steering our efforts and selecting value options.