A new product review form

Results at a glance

More reviews

One guided form replaced three different versions on iOS, Android, and Web, increasing the number of reviews by 4%

Longer reviews

Attribute sliders and category-specific prompts help shoppers organize their thoughts, resulting in 12% more long reviews

One flow instead of three

The form is live for 100% of users across all platforms, creating a consistent experience and reducing maintenance effort

Context

In online shopping, reviews provide important social proof. Our internal research showed that 60% of Lamoda users visit reviews looking for answers to specific questions. If reviews are unhelpful, shoppers hesitate, decide not to buy, or order without enough information and later return the product.

When the project began, our review forms were inconsistent. The iOS and Web versions were outdated and asked for little useful information beyond a text comment. Android had already had an attempted update, but the form was overloaded and still failed to solve the main problem: people kept writing short reviews like ‘great,’ which helped nobody.

We had three versions of the same process, creating friction for both users and engineers.

Problem

I began by analyzing Lamoda reviews and found that most consisted of one or two words. Without prompts or structure, people rarely added detail. They often commented on fit without including their own measurements, making the information hard to use. Photos were uninformative too, because the form did not explain what would be useful to show.

I also analyzed quantitative research, identifying the main review topics, the kinds of reviews shoppers found useful or distrusted, and the barriers to completing the form.

Hypothesis

If we replace the empty text field with a structured flow — attribute sliders, category-specific writing prompts, and a clear request for photos — then users will write detailed, useful reviews without abandoning the form. Quantitative research showed that people were willing to write reviews but did not know what to say. Without structure, even those who wanted to help left short, uninformative comments.

Solution

I chose to redesign the flow rather than reuse the Android solution, creating one form for all platforms. The goal was to help shoppers write a useful review quickly.

To avoid overwhelming users, I split the journey into three steps.

Step 1: a rating from 1 to 5

We start with a simple rating screen built around the product photo and large stars. The user taps a star, sees the rating animate, and is already engaged in the process.

Step 2: attributes, review, and photos

This step collects the information future shoppers will look for in reviews. I added a new attribute rating section. Instead of describing quality entirely in words, users can move sliders. For clothing, these cover sizing and fit; for cosmetics, lasting power or effect.

To avoid confronting users with an empty field, I added category-specific prompts. For clothing, we ask about the fabric; for beauty products, the texture. This helps people organize their thoughts and saves time.

I also added guidance on which photos to upload and onboarding examples.

Step 3: about the reviewer

Finally, we invite users to add their height and body measurements: bust, waist, and hips. This step is optional, and we only ask once. The system saves the information and prefills the next review, while allowing measurements to be edited at any time.

Desktop

The Web form was the final step in bringing all platforms into one system. I rebuilt the layout to make effective use of larger screens.

Results

The form is live for 100% of users on iOS, Android, and Web. Three different versions have become one consistent process that is easier to maintain.

The experiment supported the hypothesis: when the form helps people organize their thoughts, they are willing to share details. The number of reviews grew by 4%, so three steps did not become a barrier. Long reviews increased by 12%. Reviews began answering the questions shoppers actually had about sizing, fit, and fabric quality.

NMV and RRR results remained inconclusive during the test. We expect their impact to emerge gradually, as more useful reviews accumulate and influence purchase decisions.