Back Home
How I turned 428 vague reviews a month into clear, actionable insights, and boosted customer trust overnight.
Aug 23, 2021
•
7 min read
Last Updated: September 4, 2025
By late July 2021, the platform had grown fast, connecting thousands of businesses and customers. But with growth came more eyes and more pressure. Reviews became the main trust signal, directly shaping how people chose where to buy or sell.
Then the messages started rolling in. Business owners were confused and frustrated. They kept getting bad reviews, but none explained what went wrong. Some said deliveries took too long, others mentioned missing items or poor service, yet everything ended up as the same one-star rating.
A system built to create trust was now doing the opposite, and that’s when we stepped in to solve it.
The Problem
Looking closer, we found that the review model was too simple, collapsing every part of the customer journey into one star rating. This created two major issues.
Low-quality feedback: Most users gave a rating without context. A delayed delivery could receive the same rating as a missing item or poor service, with no way to know the difference.
Misaligned reviews: Roughly 8% of all reviews (about 428 out of 5.344 per month) were low ratings for very different reasons, but the lack of detail left businesses with no actionable insights.
Stalled improvements: Without clear causes, businesses couldn’t close feedback loops or fix the real problems behind dissatisfaction.
These signs showed that something was off in how feedback was being collected. We dug into the data to understand exactly how and why useful details were getting lost.
The Hidden Gap Between Ratings and Insight
Total reviews
1-3 stars reviews
No context
Nearly half of all low ratings came without context, leaving businesses unable to trace the source of dissatisfaction.
Misaligned reviews
40.6%
Low ratings for unknown reasons (Monthly)
Actionable reviews
Low-quality (no context)
The numbers confirmed what we suspected. Although reviews without context made up only 8% of total submissions, they represented nearly 40% of all low ratings between one and three stars. Almost half of the negative feedback gave no clear reason for user dissatisfaction, leaving businesses blind to what was actually wrong.
The original interface helps explain why. The original review screen only had one-star rating and an optional comment field that most people skipped. There was no way to separate issues with product, delivery, or service. In the end, businesses had plenty of ratings but very little insight.
This discovery became the turning point. With data in hand, we needed a cross-functional effort to redesign the entire review experience from the ground up.
Business ratings displayed only as a single star score
Review flow limited to one overall rating and optional comment
The Team
I worked closely with Engineering to make sure our ideas were feasible and wouldn’t break existing data structures. This partnership helped us build something that looked simple but worked flawlessly under the hood.
Together with Product Management, we kept the initiative aligned with business goals and secured space for it on the roadmap. Meanwhile, Customer Support provided crucial input. By reviewing hundreds of support tickets, we found that most complaints fell under three categories: product quality, delivery, and service. Those insights became the backbone of the new review structure.
My Role & Solution
Leading design, I shaped the new multi-factor review system around those three key areas: product quality (how good the product was), delivery experience (how quickly and reliably it arrived), and service/support (how the overall experience was handled).
The goal was to make detailed feedback effortless. I designed a structured yet lightweight review flow that guided users step by step, helping them provide more complete information without extra effort. This kept completion rates high while giving businesses the clarity they needed.
Early sketches exploring multi-factor review categories (product, delivery, service).
Paper wireframes testing how to display structured ratings and comments.
Prototype of the structured flow guiding users through detailed feedback.
Interactive prototype in Figma showing the multi-factor review flow and comment input, used to test navigation and validate usability before implementation.
Businesses now receive detailed category ratings, not just a single score.
Simplified review flow with three factors, improving clarity without adding friction.
Results & Impact
The new review flow shipped just a week after approval and quickly showed results:
With clearer data, product and operations teams could finally pinpoint what needed improvement, whether it was delivery logistics, product packaging, or customer communication.
Testimonial
The response spoke for itself.
I actually leave reviews now
2 weeks ago
LucasR87
I used to skip reviewing because it felt pointless.
Transparent and easy
3 days ago
Marcos Filho
Super simple. Leaving feedback takes seconds and businesses actually respond to it now.
Good idea, still room to grow
5 days ago
AnaPaulaTJ
Helps a lot, but sometimes I wish I could add photos. Hope they expand it.
Business Feedback:
“People leave more reviews now. They like that we actually pay attention, and it helps us look more trustworthy.”
– Sol&Neve's Owner
“Support is simpler. Instead of chasing down random one-star reviews, we know the real reason behind it.”
– Jéssica, Support Lead
“It finally gave us clarity. Before, a low rating told us nothing. Now we know if it’s about delivery, product, or service, and we can act on it. Customers trust us more, and we’ve already seen better feedback as a result.”
– Hamburgão's Owner
Read more
Let’s build something together.
