Finding a suitable card
Finding a suitable card
Challenge
Finding just the right card for someone can be tricky and time consuming - especially when there are thousands of designs to choose from.
When I first started this project, you required a little of patience to find a card. Firstly, you would select a category to take you to the products results view. If you were lucky you would get a suitable selection of cards to browse through. However, if you couldn’t find a card, you would need to go back to the previous stage and then select another category, and repeat, and repeat, etc...
This problem was compounded by the fact that the products results view only displayed nine, rather small, images at a time. To view a larger image you needed to select each product one by one, and to load more products, you needed to continuously select the 'Next page' button.
Actions
Based on my previous experience, I knew the users of the Moonpig service could broadly be divided into two groups:
- “I need any suitable card right now!”
- “I want to see all the suitable cards and choose the perfect one!”
My focus was primarily the first group of users and to target a solution that would optimise the process of finding a suitable card by removing barriers. Solving the problem for this group would also benefit the second group and not detract from their experience.
I realised during the user research process that the product database was written in a way which had not considered that a product may exist in more than one category. Each card was only suitable for one occasion (e.g. birthday) or one recipient (e.g. dad) and belong to one design group (e.g. humour). This was a problem because a card might actually be suitable for more than occasion or recipient and be classified in more than one design category.
Based on this insight, I created a new faceted taxonomy using card sorting studies, behavioural analysis and sales data. I prototyped many variants of the products results view based on the new taxonomy which I tested internally and with users - some were dynamic HTML experiences.
After validating the new taxonomy with users, I organised and managed the entire card classification project to create a faceted, searchable index. I recruited a team of people and instructed them on how to manually classify each card. It was critical for the system that this team could objectively distinguish each classification (e.g. 'cute' or 'funny').
Once we had created the index, we developed the interface and the team made a succession of product releases, iteratively adding and refining features. The impact was measured at each stage using split URL testing software to gauge the additional product value.
Outcome
The main win for the user was improving the flow by removing the dead ends. Users were able to quickly refine their products results without leaving the view.
We achieved this by helping our users find a suitable card using a combination of features.
1. A simple facet interface
enabled users to quickly refine the selection of cards without leaving the product results view
2. Larger product images
enabled users to review and compare products without leaving the product results view
3. Introducing an
infinity scroll to the product results view meant customers could quickly scan through hundreds of products with minimum effort.
For the business, there were significant benefits too.
- Conversion rate of users searching for products increased
- Conversion rate of users buying a card increased compared to the previous user flow
- Exposing the "large card" option increased the percentage in large card sales