Peloton
Helping Peloton owners leave more useful reviews
Fewer than 1% of Peloton hardware customers left a review. As feature lead, I redesigned when members were asked and how the review form worked.
Context
Engaged owners, almost no reviews
Fewer than 1% of customers left a review after buying Peloton hardware. Reviews act as social proof for people deciding whether to buy, so the gap mattered.
Other data showed that customers were highly engaged and satisfied. If people were happy and using their products, the problem was more likely to be how and when we asked.
I was the product designer and feature lead on the redesign, working in a global cross-functional team. I led the submission experience. Research, the build and the reviews page were shared with colleagues.
We set two goals: more completed reviews, and more useful, structured feedback that answers the questions buyers have.
Timing
Asking during early ownership
The old request arrived about two months after purchase. We brought it forward to whichever came first: seven days after purchase or ten completed workouts.
Review form
Rating first, personal details last
The old form was one long page. It opened with name, email and city before asking for a rating, then mixed product ratings, a recommendation score and questions about how people ride.
Before designing, we looked at review flows from Amazon, Airbnb and similar products, and read through support feedback and abandoned reviews. We then ran moderated and unmoderated research, including sessions with 10 members in the US and Canada on mobile and desktop.
The first change was to open with a star rating, an easy first step before any personal details or writing. Participants responded well to it. For returning customers, we filled in what we already knew from their account.
We then tested a segmented flow, with separate steps for the review, extra details and personal details, against a single page. Participants found the single page simpler, with no moving between sections, so that is what we built.
Ratings
Detail that helps the next buyer
Familiar 1–5 star ratings captured feedback on product quality, value, ease of use, variety of classes, and delivery and setup. We chose stars over a Likert or visual analogue scale because they are simple and familiar.
On the reviews page, each area shows an average, so buyers can compare on the same terms and the product team gets structured feedback.
Why do you ride?
We also asked members why they ride. We explored four layouts and chose the simplest, a set of multi-select tags. An accordion with “+” symbols made people expect more options to open.
Some participants first saw the question as serving Peloton’s data needs rather than their own. Once they understood it, they valued being able to say why they ride.
Photos
Making photo uploads feel safe and worth doing
Photos make reviews more credible and show the product in real homes, so we brought the upload option into the early part of the form.
Testing surfaced three hesitations. Some people were self-conscious that their workout space didn’t look like the polished images on social media. Others wanted to know how their photos would be used. Desktop users found uploading awkward.
The form encouraged real spaces over polished ones, explained clearly how images would be used and suggested uploading from a phone. We also explored saving progress so people could switch devices part-way through, but that did not ship.
Where the tips should live
We put photo tips in an optional accordion. A modal interrupted the flow and a tooltip was easy to miss.
The final form brings everything onto one page in three parts: your review and photos, additional details, then personal details.
Result
What shipped and what changed
Completed review submissions rose 50% in the first month after launch. That result belongs to the redesign as a whole, from the earlier request to the new form.
What shipped: the earlier review request, a rating-first single-page form, 1–5 ratings for each area shown as averages on the reviews page, the question on why members ride, and photo uploads with optional tips. The specification and acceptance-testing board below documents the final modules, components and states.
What I took from it
The most useful early signal was the mismatch between high engagement and almost no reviews. It pointed us at the request rather than the product.
Testing also showed how quickly a helpful question can read as data collection. Members came round once they saw the point of it, so the reason for a question needs to be clear before the question itself.
