Reviews and Ratings
Pomelo App (iOS, Android) and Website
​The Reviews and Ratings feature was introduced to increase sales by building customer trust, improving conversion rates, gathering valuable product feedback, and increasing overall engagement.
Before this feature existed, customers frequently contacted the customer support team to express dissatisfaction after receiving their orders. Common issues included sizing inconsistencies, color differences, and fabric discrepancies compared to product images. By allowing customers to leave reviews and upload photos, both positive and negative, we introduced greater transparency to the platform and helped future customers make more informed purchasing decisions.
Customer Review Requirements
The following requirements were defined for submitting a product review:
-
As a customer, I should be able to write a review title
-
As a customer, I should be able to write a review, which is required
-
As a customer, I should be able to upload up to five photos
-
As a customer, I should be able to rate the fit as loose, true to size, or tight
-
As a customer, I should be able to rate product accuracy, based on whether the product photo accurately represents what was received
-
As a customer, I should be able to rate the design and style
-
As a customer, I should be able to rate the quality
-
As a customer, I should be able to provide my height and weight for internal feedback purposes
Internal Review Moderation
To maintain content quality and reliability, an internal moderation system was required. Reviews could become misleading or unhelpful if the information provided was inaccurate or irrelevant, so moderation was essential.
The approval system requirements were defined as follows:
-
As a Pomelo team member, I should be able to view all submitted customer reviews
-
As a Pomelo team member, I should be able to approve, reject, or delete a review
-
As a Pomelo team member, I should be able to partially approve or partially reject a review based on its content
Incentives and Engagement
To encourage customers to submit reviews, we introduced a reward-based incentive. Customers who submitted a product rating would receive a 2 baht reward.
An additional concept was introduced through the Style Gallery, a section positioned higher on the product details page that showcases selected customer photos. If a customer’s photo was chosen to be featured, they would receive an additional 10 baht reward. In total, customers could earn up to 12 baht per review.
Entry Points and Access Control
The first design consideration was determining how customers could access the review flow, to maintain trust and relevance, only logged-in customers who had previously purchased the product were eligible to submit a review. This ensured that all reviews came from verified buyers and prevented low-quality or misleading submissions.
Two primary entry points were defined:
-
Access via Order History, where customers could review items they had already received
-
Access via push notifications or email prompts triggered after delivery
While these entry points covered expected review behavior, they relied on customers proactively navigating to their past orders or responding to reminders.
​
To increase participation, I introduced a secondary discovery mechanism on the product details page. A small informational prompt highlighted that reviews could earn up to 12 baht, paired with a “Review Now” call to action.
If a customer had previously purchased the product, selecting this CTA redirected them to their order history, where they could submit a review for that specific item.
This approach maintained strict eligibility requirements while expanding visibility and reminding qualified customers of the opportunity to contribute, helping drive review volume without compromising authenticity.
Write Review CTA via Order Details
Product Details minimal CTA
Design Considerations
Several key considerations needed to be included of the review submission experience:
-
Since customers could have multiple products to review, the product being reviewed was clearly highlighted to reduce confusion and help customers recall the specific item
-
To reinforce motivation, the reward structure was clearly communicated, encouraging customers to complete reviews and continue submitting more
-
Image upload guidance was essential due to quality constraints. Customers were prompted with best practices, including using good lighting, portrait orientation, and clear image quality to ensure the best results
-
Height and weight information can be sensitive, so customers were reassured that this data would only be used internally to improve sizing accuracy and would not be publicly displayed
-
As customer photos were being collected, the terms and conditions were updated and clearly surfaced so customers could understand how their images might be used across the app and website
Empty state for entering a review
Filled-in state for entering a review
Review Submission and Validation Logic
The core task of the experience was the review submission itself. Customers were given the ability to provide both qualitative and structured feedback through the following inputs:
-
Upload up to five photos
-
Write a review title
-
Write a review description
-
Rate fit, product accuracy, design and style, and quality
-
Enter height and weight for internal analysis
To ensure data completeness and consistency, validation rules were introduced based on the defined requirements.
A review description was required for submission. While the title was optional, it could not exist without supporting content, as a title without a review added no meaningful value. If a customer entered a title without a description, an inline error message was displayed to clearly communicate that a review description was required.
Similarly, all four rating dimensions were mandatory. Partial ratings would compromise the accuracy of the overall product star rating and reduce the usefulness of the feedback for future customers. If a customer attempted to submit a review with only some ratings completed, an error message indicated that the remaining ratings were required.
These validation rules ensured that every submitted review met a consistent quality threshold, supported reliable product scoring, and delivered meaningful insights for both customers and internal teams.
Error for a missing review
Error for incomplete ratings
Drop-off Mitigation and Exit Intent
To reduce abandonment during review creation, an exit-intent modal was introduced when a customer attempted to leave the review screen. This intervention was designed to address drop-off without feeling punitive or disruptive.
​
The modal clearly communicated the consequences of exiting the flow before submission. Customers were informed that canceling the review would result in losing their reward and that any information already entered would not be saved.
​
These messages were intentionally prioritized. Highlighting the missed reward reinforced the value of completing the review, while clearly stating that entered data would be discarded leveraged the time and effort the customer had already invested.
Together, these cues encouraged customers to reconsider exiting, particularly when they were already partway through the review.
This approach balanced transparency with motivation, aiming to reduce unnecessary drop-off while still allowing customers full control over their decision to exit the flow.
Exit-intent modal
Post-Submission and Review State Management
Once a review is submitted, it enters a pending state while the Pomelo team reviews the content. This review step ensures that submissions meet quality and relevance standards before being published.
​
To reinforce transparency and set clear expectations, customers are able to see which of their submitted reviews are in a pending state. Making review status visible reduces uncertainty and helps customers understand that their submission is actively being reviewed.
During the pending stage, customers are allowed to edit their review a single time. Allowing one edit provides customers with flexibility to correct genuine mistakes, such as incorrect ratings or typos, without creating an open-ended revision loop.
​
Limiting edits to a single opportunity helped reduce repeated resubmissions and operational overhead for the moderation team, while still offering customers a sense of fairness and control. This approach balanced customer experience with efficiency ensuring the moderation process remained scalable as review volume increased whilst also reducing the margin for error when granting rewards.
Example review in the pending state
Edit explanation modal
After a review has been edited once, the EDIT REVIEW CTA will no longer show
Rejected Review Handling
If a customer’s review is rejected during the approval process, they receive a follow-up email directing them to the relevant terms and conditions. This ensures customers understand why their review may not have met the required guidelines without exposing internal moderation criteria.
To avoid drawing unnecessary attention to the rejection flow and to reduce the risk of repeated low-quality submissions, customers are not allowed to resubmit a review for the same product.
​
Approved Review and Reward Feedback
Once a customer’s review is approved, the corresponding reward is immediately credited to their account. This creates a clear and timely feedback loop that reinforces trust and delivers on the incentive promised during the review submission process.
Customers are also notified through email or push notification, confirming that their review has been approved and that the reward has been issued.
​
On the order details page, the approved review is clearly surfaced alongside the average star rating. Anticipating that customers would be curious to see their published content, a “See Your Review” call to action was added. This directs customers straight to their public review on the product details page, allowing them to view their contribution in context and reinforcing a sense of ownership and participation.
Example review in the approved state
Upon clicking the CTA, they will be redirected to their review
Review Display and Information Hierarchy
Product reviews are shown directly on the product details page, with a strong focus on visibility and scannability. Above the fold, customers can immediately see the overall star rating for the product. Pressing this rating anchors users directly to the reviews section, reducing friction for customers seeking validation before purchasing.
​
Within the reviews section, the five most recent reviews are displayed by default to avoid overwhelming the page and to preserve visual balance. Customers who want deeper insight can select the “See All” call to action to view the full list of reviews.
​
At the top of the section, we present aggregated review data, including the overall star rating, fit distribution, and average ratings for product accuracy, design and style, and quality. This allows customers to quickly assess key decision factors before engaging with individual reviews. To reduce ambiguity and improve rating consistency, each rating category includes a question mark icon that customers can select to view a brief explanation of what that rating represents.
​
Fit feedback is presented as a scale rather than a star rating. This decision was driven by the nature of the data, as fit does not follow a linear one-to-five model and is more accurately described in directional terms such as too small, just right, or too big.
Using a scale makes it more intuitive for customers to understand the results in clearer. It allows customers to quickly understand sizing tendencies without misinterpreting a numerical score.
​
Below the summary, full customer reviews are displayed. To protect privacy, customer first names are obfuscated. Each review includes uploaded photos when available, the review title and description, ratings, and the date of submission. To further support fit comparison, we also reveal the size ordered by the reviewer, helping customers contextualize feedback and make more confident sizing decisions.
This structure balances high-level insights with detailed feedback, enabling customers to move from quick validation to deeper exploration based on their level of intent.
Review section on the product page
Rating tooltip example
All Reviews page