CASE STUDY
Summary
Bringing Salsify’s Insights offering to market was a top company initiative in 2017. Ecommerce is ruled by algorithms and ever-changing consumer demands, and Salsify has long argued that brands need to consistently optimize their online product listings in order to stay at the top of their game. Insights would bring those optimization recommendations directly into the Salsify platform.
I was brought onto the Insights team to help them bring MVPs of various Insights-related features into the app. Since Insights introduced reporting into the platform, the rules and patterns I defined for these projects were adopted across the product and engineering organization for all reporting work. This case study highlights some of the Insights work I did over the course of a year.
Testing Offline Deliverables
Understanding what was required of a best-in-class analytics offering could not be done in a vacuum. Our Insights team – even without a dedicated UX resource – had been a highly research-focused, customer-facing group from Day 1.
When I joined the team, they were in the process of collecting customer feedback on some Excel reports we were generating on an ad-hoc basis. This was a low-lift way of testing the value and usefulness of reports before deciding to invest the technical resources into building them in-app.
Various offline reports that aggregated product data in different ways
Defining the UI
Better bulk views of data in Salsify had been a recurring request for years. We weren’t surprised to hear that customers also wanted to consume these Insights reports in bulk.
But getting that feedback early was particularly great here. Since this was really our first case of reporting in Salsify, I’d been thinking of Insights in a bit of a vacuum, deserving of its own experience. In actuality, however, this was really no different than the other use cases for “grid view”: an Excel-like UI pattern that another team just so happened to be taking on in parallel.
Instead of designing an Insights experience from scratch, I paired up with my teammate leading the UX for grid view and shared the Insights use case. Over the coming weeks, we defined, refined and tested the bounds of this pattern together, making it even more widely applicable for use across the platform.
Adapting our existing "grid view" UI for insights
Optimizing the UI
One issue was that the grid would have no sort and filter capabilities in time for our MVP. Fortunately – contradicting the UX team’s hypothesis – we had learned from early feedback that brands would rather see their products sorted by top-performers than by grades. If we could only define the order of the products when the page loaded, we would order them by salesrank.
Another UX challenge with the grid was that it was hard to understand the content out of view. Just because the data you initially saw looked good didn’t mean that there wasn’t a minefield of errors hiding off the screen. While we worked on better summary and navigation controls, understanding that brands could most easily take action on copy and least easily take action on reviews helped us organize the columns by actionability.
Optimizing the UI
Iterating & Refining
Over various rounds of feedback, we learned that we needed to make the digestion of the data even simpler, especially for brands who could have thousands of products in a single report. Examples of feedback included:
Color coding every cell was too overwhelming. Users just wanted to see cells with errors.
Users wanted the grading criteria for each column visible as they digested the report. They didn’t want context hidden behind a tooltip.
We were also auto-generating user tasks for every poor grade found in a new report. We quickly got feedback that this made an impossible number of tasks to get through, so I worked with my UX partner to define a cohesive task creation experience in the grid view to accommodate the creation of manual tasks.
The final color palette for our first release
Expanding the Experience
Creating a home for Insights in the platform was a fun and never-ending challenge. We were constantly testing the extensibility of design decisions as we held them up against new reports and a never-ending stream of feedback.
Actionable emails, an Insights inbox and a configuration screen all borrowed heavily from existing UI patterns to start. But with increasing pressure to make Insights more visible and accessible for customers, I was able to push the boundaries of these patterns to accommodate for the best digestion of data.
The final color palette for our first release
The Solution: Digestible, urgency-inspiring data
Users depend on Salsify to cut through the noise of all of their data in the platform and tell them which activities will have the biggest impact. I held this expectation close to my heart for this project, particularly under the limitations of the MVP. From emails to configuration screens to the actual reports themselves, the data is pre-sorted and highlighted to show users their highest-priority products. While some of these UI patterns wait for additional functionality to help users further slice and dice the data, the UX biases toward clarity over breadth to laser-focus users on their most high-impact activities.
Built for Action
Heatmapping errors lets users quickly narrow down to problem areas and assign tasks with pre-filled, contextual copy
Task inbox
Subtle enhancements were made to evolve UI patterns to meet the needs of Insights, like “read/unread” download treatments for new reports
Smart, easy report configuration
Navigating through thousands of attributes can be a mess. The report configurator helps users easily parse through their data, or modify pre-configured groupings.
Merchandized email insights
Instead of typical reporting emails, Insights emails are modeled after the eye-catching, image-first templates brands would send to their customers
