Location: Stockholm, Sweden.
Project length: 4 months
Tools used: Sketch, Invision, Zeplin.
Team Members: Project manager, Front-end developer, 2 back-end developers, UX Designer(me)
Adlibris had built a good reputation over the years by being a trustworthy e-commerce website that delivered books on time. By 2015 it started diversifying its portfolio of products and adding more categories over time. By the beginning of 2018 it started having issues delivering on the promised time, but it all came down to Black Friday and Christmas season. That year they added the toys category and many parents ordered toys for their children one month in advance, expecting to be delivered in time for Christmas day.
Adlibris ended 2018 in the worst possible way. A logistical accident meant that customers had to wait weeks for their Christmas presents. Some did not receive their ordered presents before Christmas Eve, and a large customer storm raged on social media.
As a result, there were hundreds of customers who didn't get their presents on time. It hit the media and as a result orders started slowing down.
When we dug into the matter we found out the following problems:
- When an order that had passed the promised time, it would just leave the system and fall into a “limbo” until the customer would call and ask for it and then it would get prioritised.
- In the warehouse, managers and staff were unaware of any results. The only data they would have at the end of the day would be the number of items dispatched, but they were lacking a lot of data that would help them to improve the process and know where they're standing.
I went to the warehouse to interview the floor managers of the different sections and get a better insight of what were their needs. After the interview I got two conclusions:
1. They were lacking motivation as they had no goals other than doing their job as usual. They wanted to know if they were ahead of behind schedule, if they needed to put pressure on staff and congratulate them.
2. There was no communication among departments
Although a few other things didn’t show up in the tests, we knew they could be improved. One example was the low color contrast between the background and the text of the add-to-basket button. On the desktop version we removed the suggestions bar on the right side of the page (as it was just a distraction) and reorganised the different feature descriptions of the product.
I needed to identify stations to divide the process and have a structure. Initially I counted 12 but but we filtered to the 6 most clear and vital for the process: Delivery, shelving, picking, sorting, packing and loading. The next step was to know what data was available and identify the specific needs for each station.
As part of the UI, it was also needed to define a palette of colors for 3 reasons: to have consistency and be compatible with Adlibris brand, to have a clear distinction between elements, with sufficient contrast between them and a clear intention that can convey meaning in multiple ways (such as relationships between elements and degrees of hierarchy). The choice of font was Muli given it’s a san serif and has been designed to be used freely across the internet by web browsers on desktop computers, laptops and mobile devices.
The next step after the key elements were decided, was to identify and design the data needed for a successful UX. The final result is displayed and explained below:
In the delivery section it was important to know how many trucks were coming every day of the week (1), from what company, at what time (2) and how many pallets where booked for each of the trucks.
In shelving it was needed to know how many articles needed to be shelved in every station (1), how many people where working in every station (2), the total of shelved articles (3) on the day and the amount of working hours and the number of articles shelved by every individual (4).
In picking it was needed to know how many people worked in every station (1), the total articles packed (2), estimated number of peoplle that are need on the day (3) average of articles picked by every individual(4).
In sorting it was needed to know how many articles needed to be shelved in every station (1), how many people where working in every station (2), the total of shelved articles (3) on the day and the amount of working hours and the number of articles shelved by every individual (4).
In Packing it was needed to know in every station a distinction between group and single orders (1), number of articles packed on the day (2) and starting time and number of articles packed by every employee (3).
In delivery it was needed to know the total number of dispatched orders that are on the delivery date (1), the number of orders sent on the day (2), the number of trucks coming to the warehouse every day (3), how many people where working in every station (2), the total of shelved articles (3) and data about the content of each car (4).
This was my last project with Adlibris and I left before it was implemented. During the design process I had constant feedback from the project owner and the warehouse managers. The final design was a reflection of what they needed, what was achievable with the available data and the development feasibility.
