As part of my application for the Product Designer position, I received this problem as a design assignment. There are multiple reasons for this project:

• Reducing the environmental impact.
• Making the returns process easier for customers.
• Better understanding the return reasons.

I had 8 hours to come up with a solution that would porve my thinking and design process.

Honey is designed to enhance the online shopping experience by offering personalized recommendations, answering customer queries and providing style advice. This intelligent AI assistant analyzes user preferences and behaviors to suggest products that align with individual tastes, reducing the likelihood of returns.

Easy to use and seamlessly integrated into the De Bijenkorf website, Honey transforms the way customers shop, making it more intuitive and enjoyable.

The project began with an exploration phase, during which I decided my research goals would be:

1. Learn as much as possible about the users’s behaviour towards returning bought items.
2. Understand current solutions and the competitive landscape.
   
3. Determine if and how we can leverage new technology and innovation.

By allocating time to desk research and quickly putting together an online survey that I then send to (20) pals, I was looking to answer the questions:

Here's what I discovered:

• Apparel is the most returned product category.
• The main reasons for returning apparel are incorrect size or fit, followed by the product not meeting expectations.
  
• Key factors that can help the customers before purchasing: displaying detailed and accurate size information, providing multiple images, showcasing items on models of different sizes, including catwalk videos, offering information on fabric quality, and featuring user-generated content such as customer reviews and photos of real people wearing the item.
• What can help the customers post-purchasing: streamlining the return process, improving product packaging, leveraging product reviews and ratings and personalizing customer support.

From my research, I've identified some best practices that are already in place, while others are intentionally on hold - such as customer reviews, which are still up for debate because they're typically associated with mass-market items, not De Bijenkorf's image.

While exploring different ideas, I noticed AI is becoming a big deal in similar e-commerce setups, moving beyond just personalized recommendations. Could AI handle all these practices effectively? Imagine using an AI virtual shopping assistant: you could browse, get size recommendations, place orders and importantly, enjoy a hassle-free return process, just like in a physical store.

I named it Honey - inspired by the beehive concept, although the name initially felt somewhat stereotypical. Still, it stuck with me.

Given the time I had already spent in the initial phase, I decided to focus on two key moments in the journey: choosing an item and returning the item. Mapping out this journey helped me figure out which parts I wanted to highlight in my prototype.

Since De Bijenkorf prioritizes mobile, I focused on designing for mobile and didn't worry about the desktop version. I used Propolis, their design system, to create both flows in detail, using a chat-based interface.

I integrated everything that I came across during my research into Honey's capabilities as a personal shopping assistent: asking for detailed measurements (and instructions on how to measure them), size recommendations, comprehensive product information and seamless order placement and tracking. I did the same for the second flow,  returning the product.

Unfortunately, I didn't get the job - they chose another candidate with an e-commerce background. However, my proposal was very well received. It turns out that a similar tool was already in development behind the scenes.

I was of course, dissapointed, but life moves on and lessons have been learned. The most important one is that every experience is a good experience.

The second is to never hesitate to ask questions, especially in job opportunities. With hindsight, I realize I could have been more persuasive by requesting metrics from the data team instead of relying solely on the survey and internet research.

And third: I learned that 8 hours should mean 8-10 hours, not 16. I put in extra effort because I was eager for the job, but in business, time is money and I didn't manage my time as efficiently as I should have - let alone considering the balance with personal life.

To the next one!