Content Dashboard
In collaboration with my team for Liferay · 2020
1. Challenge
The Content Dashboard is a global application that helps users to manage their content. It aims to be a one-stop-place where content authors and marketers will be able to access all the contents generated by them or other users, not only for a given site, but for all the sites in a Liferay instance. It includes also a tool for auditing web content categorized under global vocabularies, and cool metrics on content performance.
Content authors need to have a single place to review at a glance all the content generated by them: filter it by different tags, sort it by date of modification or simply search for some specific content and access it quickly.
Content strategists may also need to review all the content generated by authors they’re responsible for.
Content authors & strategists need to have insights about the volume of assets published under each of the labels in order to fill the gaps or cover the cases that are not being addressed. The Content Audit Graph provides this kind of insight taking into account the classification provided by the digital marketing taxonomies.
The all content view in the content dashboard can be overwhelming, considering that all web contents of any instance will be displayed. A good set of filters will help users to easily find and audit the contents they need. Management tool bar provides filter and ordering capabilities, that will apply to both list of content and audit graph (although this epic only refers to filters applied to the list)
Why?
We thought of the content dashboard for covering mainly three use cases:
For content authors, being the main place to access all their content, in order to edit, view or check performance.
For marketers, searching for content regardless of the site or asset publisher it is contained in so it can be reused for campaign purposes.
For both of them, auditing content to assess and improve content strategy execution
That’s why:
It’s an instance level application, so content can be accessed regardless of the site /Asset Publisher it is contained in.
A good set of filters is provided, considering that all web contents of any instance will be displayed
It consists of two sections: the list of contents, and the audit graph, aka Content Audit Tool.
By now the Content Dashboard includes only web content, but as we extend asset analytics capabilities, also documents and media might be featured.
2. Proposal
How does it work?
The all-content view in the main section is a table with relevant information on each of the featured web contents, such as title, author or subtype, among others. As we’ve said earlier, it is possible to search by title, filter by multiple fields and order by title and modified date.
Depending on the permissions of the user, it is also allowed to perform different actions with the web content in each row of the table, such as viewing, editing, or getting the content’s full details on a side panel.
The Content Audit Tool in the top section contains a graph that calculates the number of existing assets for each combination of categories in the featured global vocabularies (we’ve also made some changes there… Please check this video for further information). The graph can be configured so users can select which vocabularies to display and audit the content they need.
But, what is a content audit exactly?
A Content audit is the way that content teams have to assess their content strategy execution. They run content audits on a regular basis and review all content generated according to the topic, audience and funnel stage at which it is targeted. This way they’re able to spot gaps, identify areas of interest and learn from the best performing content.
With our Content Audit Tool, content teams are able to run content audits whenever they need to.
To see contents represented in the graph, the first thing to be done is to add some categories to the OOTB Global Vocabularies, and then assign them to the web contents. You can also create your own Global Vocabularies or edit the existing ones, but be aware that just Global Vocabularies can be selected on the graph configuration.
Once you have the web content with Global categories assigned to it, the graph will show by default the Audience and Stage categories. It can be changed at any moment by clicking on Configuration on the top right of the page. Also, you can change the vocabulary displayed in each of the axes by ordering the selected vocabularies. You can watch this video to learn more.
This way content teams can run quantitative content audits in order to understand how many contents they have for each one of their defined audiences, stages, or topics (or whichever vocabularies they’ve created to categorize their contents) and assess their content strategy execution to fill in gaps or search for areas of improvement.
But the other part of content audits is evaluating the performance of contents. Ideally it is done by groups of assets - be it by category, asset type or author. We didn’t get so far for the 7.3, but we did include content performance metrics in the Content Dashboard. The new Content Performance Panel showing in display pages is also accessible from here by clicking on the Metrics option on those web contents with an associated display page. DXP has to be connected and synced to AC in order to see the metrics.
Process
The interviews
We set a series of interviews with marketers and content authors to understand the way they work and what information would be the most useful for them. Some of them were using excel files to have control over the amount of information they had.
We concluded to go for an MVP that included all the content assets and a tool to easily see if there were gaps to fill regarding content. We based our table and chart on the categorization of the content items.
Initially, the idea was to have a data dashboard with charts to see at a glance the main data.
But based on the conclusions obtained from the interviews, I created the first wireframes including all the content assets from all the instances in one place and including just one interactive chart to see the gaps, letting for next iterations include more analytical graphs.
MVP
Reduced version with a table with all contents in the product and the audit chart tool.
To make this work we needed to change how categories worked on portal, including internal categories (that are not seen by end users.).
We establish a research plan with hypotheses and goals and based on them, create an interactive demo and some tasks for users to perform and detect whether they were able or not, and where did they struggle. The goals were:
To find out where the web content title should lead
To validate the usefulness of the DPT icon
To validate the understanding of the audit graph legends
To validate if the filter and order capabilities are understandable
To validate if users can understand, and easily use sidebar panels (info, metrics)
To validate if icons improve the user experience for quick actions, such as view, metrics, info, and edit and if they understand the metaphor
To validate if they can easily find where to change the configuration of the vocabularies shown in the table and chart
To validate if they understand the link between categories and the content dashboard
To ask them for suggestions to improve the current content dashboard
The tasks:
CONTEXT
Imagine that you are a new IMDB marketer and you need to launch a new campaign to engage new users. This campaign is oriented toward users in the initial stage of the funnel. Your colleagues have told you that they previously created content that you can use to engage new users
TASKS
Check the content you may have available to reuse. How many assets do you have for this audience? How many for this audience and initial stage? How many Articles?
You want to know if the amount of content for the initial stage is average regarding other stages. How much content do you have for the intermediate stage?
Look for content in draft mode so as not to duplicate content. How many drafts do you have? What is the oldest draft content? What should happen when the content is in draft mode? Should it behave differently according to the different statuses?
Suddenly you remember that this campaign is going to be oriented just toward the users that love Sci-Fi. Can you tell me how many assets you have for them?
After talking to a colleague, she tells you that Fantasy films may also fit what you are looking for. Search for them
You love The princess bride and you want to know if there is still information available about it. How many languages was it translated into? How many categories are labeling this content? Now you realize that one category that is being used to label this content is wrong. Change it.
RESULTS AND CONCLUSIONS
Results showed an initial stopper when reaching the Content Dashboard and it was empty.
For the rest of the tasks, all of the users were able to perform the tasks.
Launching without content causes confusion
Participants don't interact with the graph configuration
Users feel more comfortable looking for information on the table
All participants think filters are simple and useful.
Users don't know what the term "Display page" means
Categorization flow is intuitive.
Participants easily interact with the side panel and the information panel.
Changes from Research:
Remove initial filter "My Content". We decided to remove a default filter implemented just to see the current user content based on these results. And the reason why they saw the table empty.
Title link should lead to the view mode of the page.
Include a tooltip on the icon to explain what is it.
Chart and Filters
Our next focus was to improve the chart and the filters.
We included:
The possibility to filter the table while clicking on the chart bars.
The option to switch the table axes.
Include filters: categories, tags, type of content, extension, sites, author, status...
After the first hypothesis, we run another usability test to clear out some doubts:
Interview for date filter
We didn't have a clear image of what were the needs of the users regarding the date filter. After performing these interviews we discovered that it wasn't so important and that they preferred to have other features before. For example download the data.
GOALS
If users would naturally click over the bars of the chart
What they expect to see when clicking on the bar of the chart (in simple filtering and in a complex situation).
If the information provided by the results bar is useful and understandable.
If they understand the difference between AND and OR operators regarding the use of different elements on the filter or elements belonging to the same filter. And if this information is valuable for them.
If there should be a visual difference in labels between AND and OR operators
CONTEXT
As a Disney + marketer you need to check how many films and TV Shows you have for Kids and Adults to make an audit of your content.
TASKS
What does the chart show you? What does the gray bar show? What does none mean here?
How many films do you have for Adults? Which ones are they?
(if not focused on the chart) [What feeling did you have when looking at the graph and why do you think you decided to go to the table to find the information?]
(can hover but not click) Do you think you can obtain this information from the chart?
(hands up) What do you expect to happen when clicking on one bar?
Click on it. What has happened? Was it what you expected?
Remove filters
Apply a new filter to only see your content. How many films do you have for adults?
# (Remove filters) Apply a filter to only see the films (this can be done using the checkboxes in the chart legend. It would be good to test it too). How many do you have for adults? What do you expect when clicking on the bar?
Undo the last filter
Add adults filter to the already existing Films (if they use the chart, explain to them that without using the chart). ----AND / OR ----- Would you click on the bar? What do you expect from clicking on the bar?
Do it. What happened? What were you expecting to happen?
(after doing both options) Which version do you like the most? Why?
RESULTS
The information in the bars is useful
Users prefer restrictive filtering behavior (AND)
In the case of offering the possibility of filtering by AND and OR, this can be done from within the "filter and order" and the choice of the operator that is being made is clearly indicated.
The union or separation of the labels in the filter result bar is not meaningful to users.
Users prefer prototype 1
The name "None" is unclear in the graph
Users understand intuitively that they can filter through checkboxes.
Changes
Taking into account these results we changed and adjusted the filter operators to be AND always that they were adding different filters and OR in case they were including more than one option for a filter.
We also changed the naming. None was updated to "No [vocabulary] specified"
Improvements
We included some sidebars to provide more information about the performance of the content and metadata.
This is a work in progress. We are still adding features to the Content Dashboard.
Document with possible improvements