However, the new platform required a complete overhaul of its administrative services to safely and accurately handle configuring Security and Data Access for enterprise companies.
Administration
The administrative services (admin) needed to grow quickly to accommodate the customer facing portion of the platform. As a result, user experience was sacrificed for quick development work in the early phases.
As we began to sign on more and more customers, the realization that administration also needs to be accessible by our customers became more immediate. I was brought on as the lead designer to not only improve the current state of administration, but to also redesign current workflows in order to guide incoming customer administrators through defining security access and configurating concepts with ease.
My Role
Lead user experience and product designer for the administration services of Visier People. Collaborated with PM and Development Managers to understand, conceptualize and design new features for Admin. I was responsible for creating research deliverables, prototyping, and the interaction and visual design of the service.
To familiarize myself with current features and workflows, I worked closely with previous developers who had worked on and implemented features for admin, and also interviewed Visier’s customer success team (who were essentially the customer administrators as Admin was not yet released publicly). Talking with the team provided lots of great insights, but also revealed many problems that needed to be resolved. The most pressing issues were:
My challenge was to design an intuitive administrative system with the aim of increasing customer adoption rate and building enough confidence in themselves to work through administration and decrease the reliance on Visier’s customer success team. To achieve this, I needed to set some goals with user needs and business requirements in mind.
When a customer works with Visier, they are sharing sensitive data about their workforce with us. Visier’s success is built upon the trust of their customers, which is why we make our customer’s security and privacy our highest priority. In order to use any part of the application, users must first have the correct security profiles assigned to them.
Data configuration is the most difficult out of the four parts in configuring a security profile.
This means that customer administrators must set up security and privacy configuration for their entire workforce before anyone can access Visier’s platform. Distinct security profiles silos sensitive information and can give users different view and access rights appropriate to their role in the organization. For example, HR employees accessing Visier People may be able to see compensation of the workforce but compensation data should not be shared to anyone else in the organization.
Setting up security can be overwhelming. Our data concepts are difficult to understand and even more difficult to configure correctly. As configuration occurs at the core data level, it quickly becomes unclear which data point is required to enable certain metrics and dimensions at the consumer level.
Customers are forced to navigate out of Administration and into the core Talent Acquisition application to check their work, all while having to remember what they just configured. This current workflow is inefficient and not sustainable for growth.
Forces administrators to navigate away from their current work.
Enables administrators to make accurate decisions while focusing on their current work.
Doing a bit of research, I discovered that Visier's platform actually has a very accurate representation of what metrics and dimensions that are available depending on the data loaded, but does not output this information anywhere. With this information, I proposed a review workflow that assists administrators as they configure security.
This would give administrators instantaneous feedback about which metrics, dimensions, and concepts are enabled or disabled. Additionally, the review process will also show what other data points are needed to enable various other data constructs (metrics, dimensions, concepts etc… )
Userflow for creating and reviewing data configuration and member filters.
Overall interactions also needs to be considered as we merge the assisting review workflow with data configuration. Taking this chance, I replaced some of the old interaction patterns with more fitting patterns to accomodate the recent changes.
The infinitely expanding tree structure was replaced with a tab structure, displaying navigation paths and providing more clarity when working with subject references. Furthermore, I also utilized the Security Review panel to provide additional learning and guidance for administrators.
Revised interactions to include security review.
Our work on Security Review also opened the door to new opportunities. With the technology we built for Security Review, we can finally accurately answer one of the most common questions we get from our customers.. "What exactly am I getting out of this product?" To address this issue, I designed a security gated workflow that allows administrators to select and ouput a file with all customer data that has been loaded on to the platform. This solves two business use cases:
I’ve mentioned before that setting up security can be quite overwhelming. Many of the tasks end up turning administrators into drones, repeating mindless tasks over and over again. But are these supposed to be mindless tasks? The answer is no, as treating them carelessly increases the chances of making mistakes, and that is something we definitely want to avoid.
So why perform a task more than once if you don’t have to? Doing a bit of digging into other enterprise administrative services, I quickly noticed something Visier is missing: the ability to create User Groups. With user groups, administrators are able to set up a couple of core security profiles and assign them to many users at once, reducing the need for setting up and assigning security profiles to users individually.
Individual management vs Group management.
However, I wanted to take this a step further. Creating and managing separate user groups can still be a lot of work, which is why I proposed that we create dynamic user groups. Working closely with the development manager, we tested the feasibility of utilizing the active data from employee records to create user groups that updated themselves.
Instead of creating a group manually for all employees who work in HR at the Vancouver office, administrators are able to select filters based on the employee record, such as the employee role as HR and employee location in Vancouver, to auto-generate a user group that matches those records.
User Groups maintained via employee record filters.
An additional benefit to dynamic groups is that they will automatically update when new data is loaded on to the platform, removing employees who are no longer working in HR, or adding new hires into HR. Do it once!
When administration was first released, Visier implemented a draft versioning system to safeguard any sensitive information and configuration until it is approved and made ready for a public release. Drafts allowed administrators to work in a private state where they can test, share and review their work before releasing it to the rest of the organization via Talent Acquisition. This could be as simple as assigning users security profiles or as complex as onboarding customer data for the first time.
Drafts are vigorously reviewed at every state to ensure no sensitive data is accidentally shared.
It is always a good idea to have a clear picture of the work you’re dealing with when handling sensitive information. Common questions administrators asks themselves include: Which drafts am I currently working on? Which drafts are under review? Which drafts can now be updated and release for the latest public version?
An administrator must always have answers to these questions otherwise they could be a liability, not only to themselves, but to their companies. However, as the platform continues to grow, more draft states are being added to patch up the workflow, causing the draft ecosystem to become more convoluted and unmanageable.
I began by having in-depth conversations with our Customer Success and Data Management teams, getting a grasp of where every person’s role is through the versioning pipeline. I targeted the customer onboarding experience, as it was not only the most complex workflow but it also covered every other scenario of the versioning system.
From the customer interactions, to cross-team interactions, my team and I discovered that there were multiple touchpoints within the workflow that were meant to be collaborative. However, because the system was broken, it was inefficient and error prone, often resulting in breaking changes or accidentally erasing days worth of work.
Stakeholders are creating their own workarounds to fix broken collaboration points.
Visier employees across multiple teams must collaborate on a single draft over the course of a couple of months when loading initial customer data. One draft is shared amongst many people with no real way to track who made the changes or why it was changed.
With collaboration being such a detrimental part of this workflow, there was actually no real way to effectively communicate and work with other stakeholders efficiently. Each draft is a siloed artifact and can only be worked on by one person at a time.
Sketch of key collaboration points between all stakeholders.
After identifying the core problems needed to be addressed, I reframed my initial goal to improve the current system, and reimagined it completely. I proposed a Project-based construct focusing on collaboration and risk management, dropping the need for multiple draft states, and effectively redefining our versioning system as Lifecycle Management.
High level flow for project proposal.
The project based approach was inspired by boards such as Trello and Asana. The idea is to move away from a developer-centric concept of versioning and rebasing and instead focus on a more casual concept that communicates a less complex mental model to non-developers.
Projects are moved between three stages. The complexity of versioning work behind the scenes.
Collaboration and revision occurs within projects.
Projects move between two states: Open and Approved. Any project can become a collaborative effort and begins when a project owner invites additional administrators and giving them the appropriate access. Work edited within a project have their own project history, keeping track of who made the edits and when. With project history, risk and conflict management becomes easier to pinpoint and resolve, allowing administrators to identify and tackle breaking changes without having to spend hours locating the issue.
We released and tested Lifecycle in it’s beta stage with key stakeholders and received overwhelmingly positive responses. Not only did it simplify the data loading process between Customer Success and customers, it also allowed our internal users to easily monitor and assist customer administrators when conflicts and errors occur.
*Some parts of the workflow are not shown due to NDA.
