Status Update: feedback on lo-fi prototypes

In order to understand how our design of a fashion/clothing recommender system would make sense to people, we interviewed 3 users, 2 females and 1 male.

When presenting our lo-fi wireframes, we asked the user to tell us what the website is about, what they can tell us about the task that would have led to our website. Two users get the idea that this is a site for recommending items that suit people’s personal needs, but relate it to e-commerce site. Interestingly, the third user said, “this is a fashion tutor site?”

All the users found the roles in the website confusing: friends, fans, fashion neighbors, follows.

All the users like the idea of “Hot or Not“, thinking that it’s an interesting idea to seduce contribution. One user even said, “oh, love it!” However, one user complained that some site with similar feature doesn’t let the user to stop it manually, which becomes annoying.

All users couldn’t differentiate between “add an item” and “upload an item“. After explaining to them, they soon got the idea. Similar stories happened to:

  • comment” (like Wall or Shoutbox) in the Profile page – confused about how it works (“Who will do the comments? Me or others?”).
  • bookmark” – think that it’s bookmarking a web page.
  • fashion neighbor” – either think it as fashion brands or people who physically live nearby.

Two users specifically expressed that they would love to have the power to customize how they want to be recommended, such as based on style or my “group”. (This user asked previously, ” Can I join a group so that I can also get a specific group’s news feed and be recommended by it?”)

In Account Setting page, one user suggest that in terms of body type, we use pictures of body shapes instead of words that describe body types because they felt it’s too subjective to say oneself is skinny or heavy.

In the search interface, one user felt it make more sense to sort by categories, such as pants, t-shirt, skirts and etc. Also, this user would like to search in “title” or “tag”.

The users’ opinions on whether we should use exact price range on items or only price symbols differ a lot. Some said we should use money symbols because we’re not actually selling outfit. Some wants to know the exact number because they felt it’s useful anyway to have it in mind.

One user admitted that she would love to see others’ outfit, but is lazy to upload hers.

These are the comments we got from 3 users. Pretty insightful, isn’t it? We’re now working on incorporating their feedback into our interface and hopefully, they will like the new one better!

Status Report

We finally had a team name: Suit Up!

Research

So far, we conducted 3 interviews and ran a survey that has 70 returned results. If you’re interested in helping us get more results, please fill out the survey questions here!

We got some insight from the survey result. About 50% responded that they seek advise for dressing up when they need to dress up for a special occasion such as a wedding, while the other half need advise when they want to know if a specific item of clothing will look good on them. In terms of whether people keep track of clothes they think of buying, about 50% answered, “no” and 19 out of 59 people use their computer to keep track of them. All the other people will either bookmark the item, circle items in the clothes catalog or keep a written list. We explicitly asked if people would be comfortable uploading their outfit to share and got comments out of it. Only one third of people would be willing to do this. However, when we asked it if they can designate select users to share, the percentage increase to over half of the people.

We also made some visual representations of our results with some graphs.

Database Diagram

We had a preliminary database diagram that will help us to give our idea a little flesh and blood.

Plugin Prototype

We also built a simple Chrome plugin prototype to help us understand how it might work and what metadata should we collect so that it makes more sense to user while they’re using our system.

Next step

Our next step is to:

  1. figure out how our recommender system would work
  2. apply users’ feedback into our lo-fi prototype