Company (product)
HashiCorp (Vault)
My role
Design lead
UXR co-lead
Product Strategy
.svg)
Timeline
4.5 months
Partners
1 PM
6 ENG
1 UXR (part-time)
%201.png)
Why I featured this project?
0-to-1
Enterprise
Complex workflows
Systems thinking
Multiple surface areas
%201.svg)
0-to-1
Enterprise
Complex workflows
Systems thinking
Multiple surface areas
Company (product)
HashiCorp (Vault)
My role
Design lead
UXR co-lead
Product Strategy
Timeline
4.5 months
Partners
1 PM
6 ENG
1 UXR (part-time)
HashiCorp is a San Francisco-based company that provides open-source enterprise software and is publicly traded. The company has over 2000 employees.
HashiCorp Vault is a B2B platform that developers use to manage sensitive data.
.png)
Millions of times daily, HashiCorp customers make use of the Vault API. To guarantee security, developers at these companies invest significant time and effort monitoring transactions by means of third-party tools.
Furthermore, hundreds of customers operate Vault on-premises, requiring even more resources and expertise for data monitoring.
.png)
I led the design of Insight, a tool used by developer and security teams to monitor and analyze Vault audit logs.
Furthermore, I designed a co-dependent service that allows on-premises Vault customers to connect their infrastructure to HCP.
.png)
I worked closely with many stakeholders to define what needs to be tracked and the technical feasibility of doing it accurately.
Increase product engagement
Increase customer satisfaction
Increase in business revenue
Every project has its evident and nuanced challenges. This one was no different.
Multiple surface areas
This project required consistent zooming out as it spanned multiple product lines (e.g., HCP Vault, Enterprise Vault, Vault CLI).
Wide range of stakeholders
I worked alongside product managers, engineers, GTM, CS and other cross-functional partners to find a balance between functionality and deadlines.
0-to-1 across the board
Outside from some basic building blocks, there were no previous designs to lean on for key aspects of the Insight experience.
Design system considerations
To gain the support of the design system team, I had to ensure that the new interactions could work for all HashiCorp products, not just Vault.
Throughout the project I conducted workshops with various groups to gain clarity on the problem space, user needs, and potential solutions.
Using the key themes generated through the workshops as inspiration, we conducted user research studies to rope in the customer perspective on Insight.
Crafting an end-to-end user journey was crucial to understanding how all the planned features would fit together before delving into pixels.
Delivering the Insight experience required working across products, Vault Enterprise, HCP and HCP Vault.
As with most enterprise offerings, Insight had a wide range of users we needed to consider, which we divided into three buckets: admins, primary end-users, and secondary end-users.
Collaborated with the product manager and engineering leads to create multiple PRDs capturing the hybrid Insight, Link, and Vault hybrid strategy.
Early in the project I conducted workshops with various groups to gain clarity on the problem space, user needs, and potential solutions.
List of key questions that were generated from the design critiques of the current filter experience.
Are large enterprise customers comfortable with connecting their on-premises infrastructure to the cloud?
Do we have the technical knowow to ingest massive amounts of log data and surface it on the cloud?
What metrics do enterprise customers care about?
How much of an appetite is there for another analytics solutions?
What will be the business model? How much are customers willing to pay?
How will users discover Insight? What will be entry points?
Myself and the UXR co-lead conducted user research studies to gather customer feedback on Insight, using the key questions generated in the workshops as a guide.
Survey based on key themes helped us understand the type of information Vault customers were interested in.
In the 1-on-1 interviews, we got into the "why" behind the surveys and also got a chance to see what DevOps and security teams were doing today regarding Vault-related dashboards.
Delivering the Insight experience required working across products multiple products.
Working across products
As with most enterprise offerings, Insight had a wide range of users we needed to consider, which we divided into three buckets: admins, primary end-users, and secondary end-users.
Workflow for each user group
Step one of the Insight journey asks the user to perform several tasks across multiple apps to connect their on-premises cluster to HCP.
Make sure that the user doesn't have to guess what they should do next.
The user may take longer to complete certain steps, and the experience should reflect this.
Step two of the Insight journey involves notifying the user of any findings through in-app alerts and email notifications.
Notify the user only when there is a significant change that requires their attention.
The notifications should provide a brief overview with just enough context.
Step three of the Insight journey involves the user diving deeper into the alerts of interest to learn more about them.
The available details should help the user decide what to do next.
Equip the user with contextual filters that enable them to cut through the noise.
Step four of the Insight journey is about giving the user valuable ways forward when they do identify alerts of concern.
Give the user clear options for proceeding with an alert of concern.
Over time, grow the available options for various alert types.
.svg)
Introducing the idea of on-premises clusters without disrupting the primary use case of HCP-managed ones slightly increased awareness.
Sharing hybrid service details and benefits to build trust with the user.
Highest converting solution that gave each Vault service equal opportunity to shine on day 0.
Although this experience was simpler on the surface, analytics showed that users completed the workflow only 48% of the time.
A late surfacing technical constraint meant that critical credential information would disappear upon page refresh.
Using a stepper to ensure users don't miss any critical steps had a positive impact on this part of the journey.
Collecting data, while highlighting key aspects of the new service with the user.
Highlighting the need to save critical data that is ephemeral in nature to avoid a poor downstream user experience.
Reminding the user what needs to be done on their end, outside of the core app experience.
A confirmation screen that provides a clear overview of user actions, tasks remaining and preview of what is come.
.png)
A low click-through rate and the inability to deep-link to specific views with Insight caused us move on from this experience.
An email notification that focuses only on key metrics and a single CTA let to a sizable jump in the CTR.
The initial version of the dashboard solution had various issues with interaction, content layout and messaging, resulting in a low engagement rate.
.png)
Abandoned this solution soon after the release as data showed a low engagement level due to various visual hierarchy issues.
Minimizing chart types and ensuring each data point had a clear CTA improved readability and usability.
The table view worked well for displaying alerts, but users encountered interaction issues when seeking more information.
Abandoned this design concept for exposing additional alert details as it caused a number of jarring interaction issues.
Offering users a two column layout to quickly scan through relevant alerts and their details increased page engagement by 13.9%.
The ability to quickly navigate the details of the available alerts improved the workflow for users responsible for remediation.
.png)
A last-minute request from Vault leadership to include the EU market for the initial launch. GDPR regulations required that users in the EU be given the option to either opt-in or opt-out of Insight.
At the time, our front-end engineering team was stretched thin, so my PM and I devised a solution that would meet the privacy requirements while also reusing front-end design components.
Step two of the Insight journey involves notifying the user of any findings through in-app alerts and email notifications.
Notify the user only when there is a significant change that requires their attention.
The notifications should provide a brief overview with just enough context.
Step three of the Insight journey involves the user diving deeper into the alerts of interest to learn more about them.
The available details should help the user decide what to do next.
Equip the user with contextual filters that enable them to cut through the noise.
Step four of the Insight journey is about giving the user valuable ways forward when they do identify alerts of concern.
Give the user clear options for proceeding with an alert of concern.
Over time, grow the available options for various alert types.
After numerous small and large revisions, countless stakeholder meetings, tradeoffs, usability tests, and design critiques, here are the results.
.png)
.png)
.png)
Insight and other hybrid services are currently in private beta.
The on-premises enterprise customers that HashiCorp is targeting are gradually opening up their infrastructure to the cloud, beginning with their development environments.
Key learnings from the project:
If you need more context and or interested in exploring my other cases studies, please get in touch.
408.335.9594 / gabeali@gmail.com / Calendly
Abandoned this approach soon after launch as the data showed that appearing below the fold caused users to miss the new feature.