VMware Security Recommendations

Identifying the problem

The VMware Carbon Black Cloud (CBC) is the company’s flagship security product. It is a B2B application designed to protect against security threats on employee-assigned devices, like laptops and mobile phones, as well as servers, USB devices, and Kubernetes containers.

Users of the CBC range from highly experienced security professionals dedicated to keeping their environment safe to less experienced teams that manage security part-time in addition to many other general IT responsibilities.

The process of tuning security settings on the CBC requires advanced technical knowledge and is often very time consuming, even for highly skilled security teams. This means that, although the CBC is a high-end security tool, if teams lack security knowledge and/or do not dedicate many resources towards maintaining their settings, they could still be vulnerable to attacks.

Given this, the team began thinking of ways we could automate the tuning process and reduce the time users spend managing their security settings.

The team

The team evolved as the initiative formed. Key team members included:

• Senior UX Designer (me!)

• Product Manager

• UX Content Writer

• UX Researcher

• Technical Product Manager

• Engineering Manager

• Engineering Architect

• Machine Learning Engineer

• 3 Full-Stack Engineers

• QA Engineer

Finding our north star

After carefully conducted research and many discussions, we identified our north star: a recommendations experience that continuously guides users in managing their security based on patterns in their environment and industry best practices. Recommendations would provide support for ongoing security settings maintenance, alert resolution management, setup guidance, and more.

In the spirit of being agile, as a first step towards our north star, we decided to focus on recommendations that improved daily management of security settings for teams of all skill levels. This would help reduce the time spent dedicated to researching new threats and would also resolve existing gaps in our customers’ security settings.

Design jams inspired ideas

I facilitated two design jams which brought together a diverse group of people from across the organization.

The ideas that came from the design jams were then used to come up with some initial sketches and workflow ideas.

Sketches became more organized and evolved into wireframes

After producing some concepts, there were some questions about the best place to see recommendations in the product. We came up with two options:

Research informed the next steps

During the first round of user research, our goal was to settle on the UX direction. The UX Content Writer, Product Manager, and I worked with a UX Researcher to build a study and create a moderator’s guide. At the end of the study, the key takeaways were:

• Participants saw value in the types of recommendations provided

• Recommendations were preferred in a single location for ease of use (Option 1)

• Participants indicated that trust was not implicit and were skeptical of implementing recommendations without detailed data-driven reasons -- we needed to first build their confidence in our recommendations

• For the same reason, participants expressed major hesitation around auto-approving recommendations, and wanted control of what was implemented in their environment

And a workflow was defined

Along with more wireframes

Then it was time to test the experience with users again

This time we had a Figma prototype that allowed users to interact with the experience. During this round of user research, our goals were to determine if we provided enough context for the user to trust and apply a recommendation, discover usability flaws, and validate the workflow.

Once again, the UX Content Writer, Product Manager, and I worked with a UX Researcher to build a study and create a moderator’s guide. At the end of this study, we identified:

• Users take careful consideration when choosing security rules and may want to further validate the recommendations on their own before applying

• Participants wanted to understand the impact a recommendation would have on their environment

• The additional data provided helped built trust in the recommendation’s value and increased adoption rates

• Users expected that the order of recommendations indicated value or impact

Feedback was translated into designs

After the second round of research, I worked very closely with the UX Content Writer to improve the designs based on our findings.

Feature development began

Once an Engineering team was assigned to the initiative, we worked very closely throughout the entire development process. I attended daily stand-ups, grooming, and regularly conducted design reviews of the completed work.

The release plan

Looking forward

As we see how users interact with the recommendations feature, we’ll be able to better identify how it should be implemented throughout the product. Long term, the goal is to improve all configuration and alert resolution management with the recommendations experience.