Disclaimer:
**Due to its nature as an internal facing website, some content and unique identifiers have been obfuscated. All information in this case study is my own and does not reflect the views of the company**
Intranets are challenging.
With lots of cooks in the kitchen, small budgets (they’re typically non revenue generating), legacy infrastructure and generally dry informational content, they can quickly become a black hole plodding along from one frustrated user to the next. Nonetheless, they are often a necessity at bigger orgs. With large employee head counts that can span business units, time zones and languages these businesses need an effective way to communicate and provide on demand access to a variety of information.
Our Task:
Our team was tasked with refreshing one such intranet whose original deployment was 18 months prior. The goal of this refresh project was to increase low user engagement numbers in 12 weeks on a shoestring budget.
My Role:
For this project, I led the initial user research, the redesign of the content publishing experience for associate users and site admins, the visual design of the updated site UI and overall project management.
In addition, I worked alongside stakeholders from each business unit’s communications team and an engineering team from the digital agency who supported the development and maintenance of the site.
Getting Started:
At the onset of the project the team was well aware of our high level goal of improving user engagement. However, we needed to gain some clarity into what that actually meant. To break it down into a more defined set of tasks, we began by conducting research in the form of user interviews and tests. Research subjects and test participants included content authors from the corporate communications team, site admins, business stakeholders and associate level users.
After organizing our feedback, we were able to identify a handful of insights, patterns and pain points that would help better define a path towards increasing engagement:
- The premise of the intranet was simple: provide associates with news and easy access to important information on demand.
- Content publishing for both content authors and associate users was a frustrating experience
- The large amount of content and lack of clear hierarchy made the site overwhelming
- Users often had trouble finding the information they were looking for.
- Content authors had no effective way to highlight the most important news
- Expansive taxonomy made content scattered and disorganized
- An abundance of features and functionality was either underutilized or redundant
We translated these takeaways into the following goals that would drive our design strategy:
Refocus attention to the site’s content
Make creating, sharing and consuming content easier and more efficient
Clearly defined content types
Use minimalism to promote organization and simplicity
From Point A to Point B
Now that our path to increased user engagement was a bit more defined, the team jumped into the design phase of our project.
Refocusing attention to the site’s content:
To allow published content to be the focal point of the site, the team took an image heavy approach to content display and moved to a lighter UI and sans-serif typeface that would further highlight content and improve readability. These UI changes also included a switch from multiple accent colors for varying content types to a single primary color that followed the org’s styleguide.
Make creating, sharing and consuming content easier and more efficient:
Prior to the redesign, publishing content on the site was a frustrating task for associates as well as the content author’s on the communications teams and the taxonomy that defined and organized content types was littered with redundant and little used terms. One of the biggest goals of the project was completely re-imagining the way content was created, shared and consumed for all user types.
For associate users and associate created content, this meant designing an experience that allowed them to create from a single source regardless of the type of content they wanted to post. We accomplished this my implementing a single share component that allowed users to create from anywhere on site quickly and easily without having to define tags and terms for their content.
When working within the confines of a CMS, much of the attention is pointed towards the end user and front end of the site. We thought this was large oversight since our real power users were the content authors and site admins that were working from the publishing dashboard day in and day out.
With with in mind, we made it a point to sit down with our content authors and publishers to better understand their current workflow and pain points so we could create an experience that allowed them to focus more on creating beautiful, engaging content and less on the logistics of displaying content on the site in a meaningful way.
Once we took a deeper dive into their day to day, we were able to outline some low hanging issues that would fit into our project timeline and budget while having a large impact on their publishing experience:
Content authors were using a WYSIWYG editor to create their articles but did not have a definitive set of guidelines for styling their posts and were using a variety of text sizes/weights/styles/spacing when creating content.
To take the guesswork out of publishing and create consistency, we restricted the content editor options so it fit within our site style guidelines.
Authors had to manually publish their posts at the desired time of posting. This meant logging in at all hours of the day/night depending on what needed to be published and made it difficult to maintain an organized content calendar.
We designed content scheduling functionality that allowed authors to create drafts and set publish/expiration dates for their stories and articles in advance. This radically changed their current publishing workflow by allowing our content authors to work more collaboratively and better manage the scheduling of their workload.
Our Content authors had little control over when content was displayed on site. The current implementation allowed authors to post to a single static hero section on the homepage or in an article feed that mixed associate and admin created content. The feed was able to be sorted by content type and recency, but after diving into the data, our team found that the sort functionality largely went unused. This meant that important admin created content was often quickly hidden by associate generated posts, pictures and videos.
To combat this issue, the team created individual components that would strictly display specific content types. This was done in parallel with the redesign of our taxonomy structure. We also gave content authors and admins the ability to select where their content was visible on site to provide more control over where and how their content was consumed.
The last part of our content redesign was improving the way in which content was consumed on site. We had already provided consistency by defining text hierarchy and vertical rhythm with the restrictions to our content editor. This immediately improved the organization and readability of our text based content.
To round things out we turned our attention to our interior article pages and updated their layout to be more focused on content. The existing layout was a three columns (20 – 60 – 20) with the left gutter occupied by the site navigation and the right gutter housing a trending article/social share widget. The new designs utilized a 2 column 20 – 80 layout that removed the right gutter and outdated widget completely. This allowed us to provide more whitespace around our posts and articles and remove the distraction of the trending article/social share widget.
Clearly Defined Content Types:
When we began the refresh project, the intranet site utilized a large taxonomy to define different content types.
These terms included:
- Articles
- Posts
- News
- Images
- Video
- Blogs
- Chats
- Slideshows
- Presentations
- Notifications
- Ads
All users were required to select a content type to publish as well as tags to further categorize content based on business unit and location. Over time, this free form process resulted in a wide range of tagged content that seemed disconnected and chaotic. To make matters worse, some terms were only used for specific widgets that utilized the taxonomy to dynamically drive content display but had been suppressed over time. This meant that if tagged incorrectly, some site content would never be visible on site.
Right away the team realized that we needed to consolidate and reorganize these terms into a simple more logical solution. After extensive discussion, we decided to move forward with a user based approach to our content organization. This would allow us to cut down on the number of terms used in our taxonomy and provide more control over where on site content was displayed. The new content type breakdown looked like this:
Company News (admin/content author):
- Articles
- News
- Notifications
Associate Voices (associate users):
- Posts
Since a large number of content types were either deprecated or rarely used, we were able to drastically cut down on our terms and because content was now tied to user types, it also gave us the ability to design componentry that would only display content created by specific user types. This provided improved content hierarchy and control over where and how long content was displayed on site. For associate created content, this also enabled us to group different media types into a single taxonomy and removed the need to the user to define the content type they were posting during their content creation flow.
Using minimalism to promote organization and simplicity:
One of the biggest takeaways we received from our initial user research was the notion that the existing layout of the intranet was overwhelming, especially on the home page. While our updates to the sites UI and consolidation of content types went a long way towards organizing the chaos, the team still felt there was more room for improvement. Namely, sunsetting deprecated and little used components/functionality, reimagining the display of our navigation and improving search accuracy.
Trimming the fat:
Because of its limited usage, we elected to completely remove the trending story section of the homepage which was a component listing articles with the most likes/views living in the right gutter. We also removed all social indicators (likes/views) from article thumbnails. Instead opting to only display those metrics on the actual content pages. Additionally, the team determined that the “add space” section of the homepage that displayed rarely used promotional CTAs could be reproduced using our new Company News component and was cut.
Calming the super nav:
Much of our user angst revolved around the somewhat confusing navigation. The current implementation utilized an always expanded super nav whose categorization had become unclear and busy with redundant content over time. We knew we were confined to our existing template with left navigation so our objective was to reorganize the info architecture in a more logical way, remove any of the dead weight and develop a design that gave the impression of order and structure while remaining in the left page gutter. The team eventually landed on an accordion style navigation with parents, children and sub-children that allowed only one menu section to be expanded at a time.
Before:
After:
Moving Into Development:
With our approved designs and functionality docs in hand, we began to engage the agency engineering team to finalize a development schedule. With input from the agency team, we decided on a hybrid waterfall methodology. Each of our tasks were prioritized and broken up into four development sprints. Each sprint would be followed by a short QA window and deployment phase. Once all items were approved by the team in our production environment, the next dev sprint began. Tasks were grouped together by their front end impact on the site with items having the highest impact on the visual site design placed in the later stage sprints and twice weekly standup meetings with the agency lead were held throughout the development phase of the project.
While there were some bumps along the way, the final sprint was deployed to production on schedule in Feb. 2017
Results:
Despite our aggressive timeline, the project was completed on schedule and on budget.
Six months after the launch of the refreshed site, user engagement increased almost 50% (metrics used to quantify engagement omitted due to confidentiality)
The team was forced to cut some desired changes due to time and budgetary concerns (this update revolved around additional enhancements to the content author publishing workflow and user BU/Location customization)
Reflection:
Looking back on the project, there are plenty of areas where we did well, fell short and can do better next time:
What went right:
The team felt that despite the limited time we had to dedicate to a research phase, we were able to accurately define the most important issues users had with the existing implementation and improve the experience in areas that would have the highest impact with the user base. We felt that despite having to cut a few enhancements due to time and budget, completing the project within our aggressive timeline was something we were all proud of.
Where we fell short:
Miscommunication with our engineering team caused a lack of understanding of how the content type taxonomy was currently structured and how SSO was integrated into the site.
The taxonomy issue forced us to make revisions to how associate users shared content. This initial desire was to enable them to share any supported content type (text/image/video) from a single place. We were able to present a single component for sharing associate generated content but were unable were forced to utilize a tabbed layout where they were forced to select the medium they were sharing.
The SSO issue forced the team to cut some of the BU and location specific enhancements we had planned. Because the intranet site relied on user data passed through SSO and did not store user data on site, we were unable to set default content displays based on user info since the SSO data accompanying the user login commonly didn’t match the location and BU taxonomy structure implemented on site.
Looking back, the team also felt that we cut corners when it came to user testing and iteration. While we were lucky to get the experience close to right the first time around, we would have liked to had the chance to test and iterate a little bit more along the way. We also felt there is always time for user testing so using the excuse that we didn’t have time is a cop out.
What we can do better:
Getting the engineering team involved earlier in the process. This could have helped prevent us from moving forward with designs that could not be implemented due to the CMS’ current limitations. Since development was handled by an outside agency we had to be conscious of our development hours in order to stay under budget but 4-6 hours during the project’s design phase could have uncovered our taxonomy issue much earlier and stopped us from spending time designing some of the location and BU specific features we had to cut due to SSO issues.
