Product Design, User Research, UX, UI
NewsCred is a content marketing platform used by some of the biggest brands to manage, publish, and measure the ROI of their content marketing campaigns.
The biggest problem we were tackling as a company in Q4 2015/Q1 2016 was customer churn. By performing a retention analysis with the product team, we hypothesized three customer behaviors that correlated to higher rates of retention.
More than 2 average unique daily logins
Used more than 4 products on the platform
Published content from at least 50% of sources each month.
During each quarter, product, design and engineering self-organized into small autonomous squads to tackle customer problems. My squad decided to focus on improving the core experience of the "Content Explorer," as the product was core to the NewsCred business, but had not been improved or touched upon in years.
Before the Redesign
Challenge: Licensed content is core to NewsCred's business, but the user experience of discovering and publishing licensed articles/images was cumbersome and clunky, resulting in a high churn rate. We believed by making the mundane task of searching for and publishing content as frictionless as possible, we would increase retention.
user research & ux diagnostic
I sat down with editors and curators to observe their workflow and understand how they accomplished their goals using the platform today. Based on these sessions, I was able to identify key pain points within their user journey. My squad and I also wrote user scenarios based on these job roles to perform a UX Diagnostic. This helped create empathy among engineers and allowed us to converge on key areas in the user experience that needed to be improved.
Curators had to be trained to use Boolean search, which was not easy to grasp, requiring overhead from internal teams. Because of this, curators would "save" popular searches so they could easily access the same type of search later.
Curators liked to filter their search by different sources but this was difficult to do quickly via the current UI, where the user had to type in the name of the source in order to create a facet. Curators also did not remember all the names of the sources they had access to.
Curators also "favorited" articles to create a backlog of content to use later in their workflows, but there was no way to organize these articles in a helpful way. For example, if a team had multiple curators in a workflow, there was no way to denote who saved which article to publish later. We found users were organizing this type of information off-platform in a spreadsheet.
mini design sprint
During discovery, I organized a Design Studio with cross-team collaboration from other departments at the company to help us ideate and converge on different ideas to prototype and test.
Some tactical ideas we converged on include
Move to natural language search rather than Boolean search
Allow users to easily save and share their searches with different team members
Redesign filters for ease of use and move to a sidebar pattern in order to free up vertical screen space for search results
Allow users to save and organize articles by lists or tags
Improve the content preview to better optimize for screen space
Several ideas were wireframed and prototyped. I conducted usability tests throughout the quarter 1-2 sprints ahead of the development cycle. I also worked closely with our tech lead to run user tests on light-weight web prototypes to learn how switching to natural language search would affect search quality.
Key insights which helped shaped our first release
As we continued testing natural language search, curators were only filtering results by source and date.
Curators need to scan search results very quickly and were utilizing the headline and source to determine whether they would read an article or not. We later removed bi-lines to increase vertical real estate and display more search results. Moving the thumbnail image to the right also improved scannability.
Users also preferred tagging over lists as a system of organization, since tags were already being used on other parts of the platform to organize assets.
The original design used a two pane layout which showed a preview of the article in a right pane. In our initial research, we noted that this layout made it hard for curators to read content quickly. We initially tested full-screen previews but curators were disoriented after closing an article. They had a hard time picking up where they left off in the search results. Thus we settled on a modal overlay that could easily be closed.
Retained 15 goal clients which represented millions in revenue
Increased average usage of content sources from 24% to 35% and created a scalable design that accommodated the business moving to an a la carte content model
Increased NPS score from 36 to 45
Decreased time curators spent in workflow searching for the right content by 12%