I led the user experience for defining a powerful and flexible content sharing framework.
About Sumo Logic
Sumo Logic is an industry leading cloud based log management tool. It's a swiss army knife for developers and ops personnel to monitor and troubleshoot their applications.
Organizations use Sumo Logic to monitor their applications and troubleshoot outages, security intrusions etc. Our users run queries to analyze their logs, analyze their metrics data, and visualize it all in dashboards.
The key motivator for this project was the highest requested customer feature in Sumo Logic - a better content collaboration framework. The sales team reported that key customers were at the risk of churning because of poor collaboration capabilities.
Highest requested customer feature
Our customers had 3 main pain points that forced them to use alternative tools -
"Sharing search queries with my teammates is painful. I save my queries in Google Docs/Evernote/tool of choice."
"I want multiple people in my team to be able to edit a shared dashboard."
"I want more control on the data my teams can access."
Users were creating shared Sumo Logic accounts or using collaborative online tools to share links to dashboards and search queries. This had a severe impact on their time to troubleshoot a problem, because different members of a team could not update a shared dashboard or search query during a "firefighting" operation.
Moreover, novice and infrequent users were often stuck in Sumo Logic because they couldn't find content created by expert users or other members of their team. This increased their frustration with the product and was reflected in their low NPS.
I was the lead UX designer on this project. I worked closely with Product Management and Engineering to define the success criteria for this project, ensure its successful Beta release, and iterate continuously on user feedback.
I took the results from a collaborative brainstorming session to define the information architecture and core user flows. I created prototypes to explore different solutions, assisted in user research to validate the solutions, and iterated on the feedback.
This project is in the process of being implemented and released to beta customers.
In user testing sessions, users have rated the new framework an average of 6/7 for ease of use and appeal.
The design process has been an effective sales tool. The sales team has used the design progress to upgrade existing customers because this was the highest requested feature on our customer portal and a competitive differentiator.
24/29 customers we user tested this with signed up for the Beta release. This showed their eagerness to adopt this new sharing model as designed.
The Design Process
Understanding the problem
I worked with a Product Manager to understand the customer pain points in depth. We scoured through comments in the customer portal, the NPS comments around this issue, and support tickets. We spoke to a breadth of users from admins in large organizations to developers in 15 person startups.
We uncovered several core customer needs:
Users want to group shared content by context. For example, by project or application or team
Users want to have team ownership of content i.e. the ability for multiple users to edit content
Expert users want to help their team mates discover valuable content
The current sharing model was severely broken. We discovered that our users were using innovative methods to circumvent the problem.
They saved raw queries in a Google Doc and shared it with their team.
They saved links to Dashboards in a shared Wiki. It was specially useful to help onboard new users in the team.
A team would create a shared account so multiple users could login with the same email and edit saved search queries as needed.
Some users, mostly novice users, had no idea there was a library where they could go see what other users had published.
Old content library
Adapting collaboration for Sumo Logic
Before exploring collaborative tools, I wanted to define a set of design principles to make sure our final collaboration model adapts to Sumo Logic's use case. I based the design principles on the initial user research we had done:
Optimize for visibility of shared content
Most users we spoke to were unaware of the shared content in their org. They needed more visibility into what was being shared by their team mates
Don't let administration get in the way of collaboration
We wanted to make it easy for the administrators to organize everyone's content but not at the expense of quick sharing
For example, if a user wants to share a dashboard with their team, we wanted that experience to be as quick as possible and eliminate any need to organize that content.
Adapt to different workflows
All users we spoke to, regarding how they currently manage their files, had unique workflows. Some users preferred organizing files neatly in folders, some depended heavily on the recently accessed list, some were heavy search users. We wanted our framework to adapt to these different workflows.
Analogous product research
I started with exploring online collaboration tools that have built entire businesses around collaboration e.g. Dropbox, Google Drive, Box, Sharepoint, and Egnyte. Each product was unique because it emphasized different aspects of collaboration:
Google Drive is heavily focused on ease of sharing and collaboration. They make it really easy to share content as the expense of organizing content. It's easy for users to create new documents and share it with their colleagues but shared content often gets lost in a big list of shared content. Users have to rely on using search for finding content. Google Drive has an excellent search capabiilty with advanced filtering to help users narrow down to their search.
Box on the other hand, like Sharepoint and Egnyte, is geared towards organization and better top down administrative control. Users can share content with each other only by placing them in shared folders. This model encourages end users to create folders and share it each toher.
Egnyte takes this one step further by creating strict demarcations between private content and shared content. Administrators have to create shared spaces (aka folders) in the root level organization folder, and create subfolders within them, and give access to the right people on those folders. Users can then add their content to those folders to collaboarate with their teams. This adds a lot of admin over head but creates a logical and easy to navigate hierarchical structure.
How does a user share content?
In the old model, users could only publish content to the entire organization. However, our research clearly showed that they cared more about sharing content with their teams.
My initial flows focused on understanding how users would share content from within Sumo Logic.
They could create folders, share it out, and add content to it
They could share individual content items from within the library
They could share searches and dashboards from the search or dashboard pages
When saving searches or dashboards, they could place it in a shared folder
Flows for different sharing paths
What can users do with shared content? (Access control)
In user interviews, we discovered that there were 2 competing user needs
Administrators in large organizations wanted granular control over sharing.
Smaller teams did not care as much about restricting access to content.
There was a strong push to cater to large paying customers and make permissions very granular. The problem with this approach was that it made sharing objects complex and confusing.
I made the case that although the admins and economic buyers would love to have the extra control, the majority of our users will be using a simple and open sharing model.
Hence, we adopted a guiding principle for addressing both these cases
Make sharing easy for most users.
Make granular permissions possible for the advanced users.
Simplified sharing options for majority users
Administrators and security experts who wanted more control over sharing could access the "Advanced View". That let them control if users could share the item with other users.
Advanced sharing for administrators
How does a user discover shared content?
Helping users discover useful shared content was the biggest challenge of this project. Enabling users to share quickly and easily was only one half of the puzzle. We had to help users find the content shared with them.
Some of our expert users were already sharing a lot of content but the old model had its drawbacks. To find content in the library, users had to know exactly who shared the content and navigate to their folder.
Old content library structure
As a workaround,
they started using shared team accounts so content relevant to a team would show up under that folder. This was against Sumo Logic policy but we did not enforce it strictly.
they kept a list of useful dashboards and searches in shared google documents
Hence, for design explorations, I had 3 main goals
Help expert users create shared team spaces
Help novice users find content relevant to their team or project
Help novice users discover new shared content
My initial proposal focused on helping users build collaborative team spaces. The 2 highlights of this new content library framework were:
The "personal" and "org" sections from the old library were merged under All Content
Shared content was no longer grouped by user name
The reasoning behind this view was
Users could create folder around projects or teams. Other team members could add to this collective pool and novice uses could treat it as a wiki.
User could see all the content in their organization thereby letting them discover content they might have missed out on in the grouped user view.
Users accustomed to the old view could still search for content created by specific users via filtering.
A consolidated view
To ensure this view was not overwhelming, I introduced the concept of tag based filtering. Users could filter this view using filters like "Created By: User X" or "Content Type: Dashboards".
Search filters explorations
Testing this model
We wanted to test this hypothesis. I worked with engineering and product management to ship this to internal users so they could play with it and tell us their experience using it.
Our primary mode of feedback was through a Slack channel. I also conducted one on one interviews with internal users, primarily developers, to get a first hand view of their interactions.
We quickly learnt that the new model was not working out for internal users. Many user's first time reactions to the new view were:
"This is overwhelming, I can't find stuff I've created"
"I don't know where my stuff is"
User feedback in Slack
Although users were pleasantly surprised by the new content they discovered, they overwhelming despised not have a safe "my space".
I gathered some valuable insights from these observations:
Users were accustomed to a "personal" view from the old library which they viewed as "all my stuff". That was a safe space for them unadulterated by any one else's content.
We need a robust strategy around transitioning existing users to the new experience.
Filtering for content was not an intuitive view for users
For example filtering by "Created By Me" showed users a flat list of their content i.e. content within folders was pulled out in this view, and users saw duplicated content. In a filtered view, the same item would show up inside a folder and outside it. This was confusing for users.
Pivoting to a a new model
To give users a safe "my space", I decided to borrow the Google Drive model. I introduced the concept of My Content, very similar to Google's My Drive.
The idea was that everything users create is added to My Content. They could also choose to add shared content, that they thought was valuable for them, to this space.
I also picked the Google Drive model because most customers we spoke to were aware of the Drive model and had been using it actively.
The Google Drive model
Testing the pivot
We were strapped for time at this point. I teamed up with the product manager to conduct quick user validation with 5 internal users and 3 customers for this model.
We discovered that although Google Drive was a familiar collaboration tool, there were severe drawbacks with it. Specifically for the Sumo Logic adaptation of it -
Users did not intuitively understand the concept of the "My Content" space.
Different users could organize the same shared item in their "My Content" as per their mental model. Hence, users did not intuitively grasp where an item actually lived.
They did not realize that reorganizing items in the "My Content" view them around had repercussions for users who shared that item with them.
The Google Drive model, although familiar, wasn't understood very well by users.
The Single Hierarchy Structure
To address the issues of content overwhelm and a confusing "My Content" model, I proposed a simple single hierarchy structure
Each user would get a Personal folder where they can save stuff they have created
They would see a single hierarchical view of everything that is shared with them
They will be encouraged to favorite content they really care about, and their Personal folder will be favorited by default
The single hierarchy structure
The final library structure
We are in the process of implementing this library structure.
We tested the prototype of the Single Hierarchy Structure with 6 internal users. It has proven to be the easiest to understand of all the models so far.
The Sales team has been using the prototypes to upgrade existing customers.
How do we measure success?
As we roll out the implementation to beta customers, I created a list of important questions to track to ensure that the new content sharing model is a success:
I worked with the product manager to define the success criteria for the new content library structure. The primary questions we want answered are:
Have users stopped using shared Sumo Logic accounts to share content?
Are users creating and sharing contextual folders to store shared content?
Are they using filters and favorites to quickly access content they most care about?