Analyzing Data and Detecting Signals
eBay Inc. is The World’s Online Marketplace®, with its website, eBay.com providing online auction and shopping capabilities to 233 million customers worldwide. SAP assisted eBay to identify a more efficient way of detecting and communicating error signals, using the power of SAP HANA, to allow for faster resolution of errors. Signal detection is crucial to the core business of eBay.
eBay’s marketing analysts are responsible for analyzing online shopping patterns for specific product categories on their site, for example, high-tech mobile phones. Each analyst looks for trends to predict buying patterns, reviews the markets, and upcoming and past events. They also look for anomalies in online trends, and are responsible for identifying the cause of these anomalies for their specific product category.
eBay were using Excel charts to categorize the trend data and emailing these charts between the geographically dispersed teams to analyze the data. This process was not scalable due to the number of people needing to work on the data. It could also lead to errors if anomalies were not identified quickly. The Design & Co-Innovation Center (DCC) from SAP was tasked with designing a solution that leverages SAP HANA’s pattern detection capabilities to enhance eBay’s signal detection.
Delivering Innovative and New Designs
From the Design Thinking workshop that was conducted to understand eBay’s needs, the DCC learned that the analysts were spending up to one month identifying the cause of errors. With this in mind, we created design concepts for how the signal detection could be handled more quickly and efficiently and reviewed them with eBay. Using a new dial-based navigation to visually represent eBay’s thousands of product categories, we redesigned the process of navigating through eBay’s data. Instead of using thousands of unclear icons to represent the categories and records, the dial was designed to be partially visible and partially invisible so as to maximize the display on the screen. The dial navigates through the different levels of data, displaying statuses at each level. Users can drill through the levels to find the information they need. To enable the analysts not only to navigate through the levels of data, but also to identify any anomalies at each level, we iterated on designs to come up with a way of flagging those anomalies clearly on screen.
Another challenge was how to display the timeline or history of tasks, such as analysts defining, reviewing, and editing signals. Timelines are commonly rendered on a horizontal access, going from A to B. This means that you are limited by screen size and need to scroll from left to right to display all the details.
“Making smart people brilliant”
By focusing on eBay’s core requirements, which were easier analysis of data and faster detection and notification of anomalies, the DCC delivered a number of tools to support them. The dial navigation, a new design paradigm, allows users to visualize information in a simpler way. Users can glide through categories and pick the ones they need like a news reel (this can be done with one hand on a tablet device). This also helps to minimize clicks and interactions.
To quickly identify anomalies in the data, we delivered a heat map that is displayed in tile format with notification of all signals in real-time. Any anomalies jump out at you right away. There’s no need to search and you can drill into any signal to see more details. To display task history, we came up with a new way of rendering information in a spiral design, based on the Fibonacci Sequence. This quadruples the real estate of the screen, accommodating more information. You can plot data of an entire year on one screen, enabling more efficient pattern finding. For all these deliverables, we ensured that we applied eBay’s branding to customize the tools to their needs.
The new tools have had a considerable impact on the job of the marketing analysts. What previously took up to one month can now be done in a single day. It also provides them a holistic view of the data so they can make informed decisions having the key data right at their fingertips. Other analysts can also contribute with recommendations to help diagnose problems and bring transparency, resulting in more accurate decision making. The tools are already in use by the customer and have moved to a custom development project. We have received positive feedback from account executors, who act as the interface between SAP and eBay.
With the support of the SAP UX Design team, we completed a successful project for eBay signal detection. The design team produced innovative and creative UI designs and provided valuable user experience insights for the project. The application that was implemented based on their UI designs received extremely positive feedback from the customer.– Frank Wu, Project Manager, eBay HULK Proof of Concept team for Strategic Customer Engagement, SAP Labs