DUA Content Dashboard


This document outlines the metrics found in the DUA Content dashboard, the context for each metric, and why these metrics matter to the intended audience. This document serves as a history of the changes made, and describes the "trail map" of future enhancements and fixes, if applicable.
The dashboard contains a mix of data sources including:
    Public Traffic view of the Google Analytics property
    All Data view of the cross-domain Google Analytics property
    Feedback data pulled from the database
    Google Search Console
    Search.mass.gov cross-domain property in Google Analytics
The goal of the dashboard is to help the Customer Experience (CX) team better understand how their changes to specific Department of Unemployment Assistance (DUA) content has on the customer journey, as well as how it effects the broader DUA organization on mass.gov.

Dashboard audience

We produced this dashboard for the CX team. We worked with Mimi, Fiona, and Dan. They will be using this dashboard to monitor content metrics across the DUA organization.

Data sources

There's at least 2 spreadsheets behind this data at the time of writing. One of them is this "November DUA Feedback," which actually has all the data needed for pages 1 and 8. There's another, "DUA_nov_dec_feedback.csv - DUA_nov_dec_feedback.csv," which we no longer have.
The query to pull feedback is:
SELECT hit_date, node_id, node_path, unique_pageviews, total_yes, total_no FROM analytics.daily_health_components where hit_date between '2020-12-01' and '2021-03-01' and node_org = 'Department of Unemployment Assistance' and unique_pageviews > 0;
You'll want to modify the date for this query and append it to the spreadsheet. Note that current settings forbid you from retrieving more than 10,000 records via Superset, so you may want to wrap the entire query in SELECT COUNT(*) FROM (paste in query) dua_query; and make sure it's under that limit. Alternatively, you can have somebody on the web analytics team pull it from the database.

Metrics and context

Below, we outline the metrics of each page of the dashboard along with context for why they matter to the audience.

Summary page

This page contains scorecards for the success metrics picked by the CX team, as well as descriptive metrics key to monitoring the performance of DUA content. They include a mix of data from mass.gov as well as cross-domain data, as denoted by section titles.
This page is fairly complicated, and includes a mix of calculated fields, blended data, data sources, etc. A couple of highlights:
    The CX team wanted to define "click through rate" as some "click" events but not others. To make this work, we created a blended data set that included only the preferred events.
    The feedback is exported from the web analytics database and lives in this spreadsheet
    Many of these scorecards are about all of Mass.gov, but a few are filtered for just DUA content
    Date range (defaults to the last full month)
    Language filter, which affects 2 scorecards ("All UI Online Form Submits (GA)" and "Unemployment Form Submits (GA)")

Traffic Overview


    Date range
    A customer filter we built that lets you filter for targeted DUA Content node IDs (selected by Fiona), all DUA content (any page with an Organization of "department of unemployment assistance"), or by all other mass.gov content not within DUA or the selected node IDs.


This pie chart shows the top browser languages set by users for the pages controlled by the DUA Filter located on the left hand side of the page.
The other metrics are typical descriptive metrics that CX can use to monitor traffic volume, bounce rate, time on page, etc.

Search data

This page shows data about internal searches and traffic from Google. The Google data is filtered for landing pages containing "unemployment" or "pua."
In both cases, you're able to look at which queries people entered, but you have to do so a bit differently. For Google, filter by "query." For internal search data, look at "previous page path," where the query parameter tells you what people searched for (e.g. "search.mass.gov?q=pua".)

Source breakdown

Top sources broken down by channel

The tree map shows the top channel groupings (organic search, social, direct, etc.) for DUA pages and the top sources within them (google, bing, unionline.detma.org).
The table below shows the same information as the tree map, but with additional metrics of sessions, pages / sessions, and exits.


    Date range (defaults to last full month)
    Channel grouping

General Traffic Flow

Where users come from to view DUA & transactional pages

This table shows how people got to pages on mass.gov, unonline.detma,org, and unemployment.mass.gov. It includes the hostname, source / medium, previous page path, page, and metrics of pageviews and unique pageviews. You can also filter by any of the dimensions except previous page path.
The intention with this table is to allow the CX team to further investigate questions about how people arrive at DUA content or DUA platforms, and how many people are doing so. This is a table that is meant to be interacted with via the filters to find interesting insights to investigate further.

Mass.gov pageviews that preceded users going to transactional sites

This section contains 2 scorecards showing how many pageviews on mass.gov resulted in people going to unemployment.mass.gov and uionline.detma.org.
The table to the right of the scorecards is a detailed breakdown of the scorecard information showing the previous page path on mass.gov, the page they went to for either transactional app, as well as the number of pageviews and unique pageviews.
The table allows the CX team to explore the user path for specific pages on either mass.gov or the transactions apps. Again, this table is meant to be interacted with and not just viewed.
Note: the filters only affect the table. They do not affect the scorecards.

PUA mass.gov pages

Because we do not have cross-domain tracking on the ui-cares-act.mass.gov site, we are not able to provide pageview data for pageviews on that site. However, we can show the clicks to this website from mass.gov.
The scorecard shows the number of clicks on the ui-cares-act.mass.gov outbound link from PUA pages on mass.gov.
The table shows the scorecard data broken out with source, previous page path, page (on mass.gov). and pageviews and unique pageviews. You can filter it by page and previous page

Landing and exit page performance

This page shows what page visitors landed on and which page they exited on. For each, a line chart summarizes the total number of sessions by day, and a table breaks out totals by page.
    Date (defaults to last full month)
    Page type
    Landing page
Understanding where people enter and where people exit mass.gov can help the CX team better understand the general paths people follow and if the right content is being presented to the user at the right step in the journey.

Topic and service page performance

Topic and service page click through rate (CTR)
This page provides additional detail for the "click through rate" key performance metric. In the topmost table, you can see CTR for each topic or service page, along with the components of that metric, unique pageviews and events. (Events here is a subset of all events--only the ones that the CX team wants included).
If you like, you can filter this table for other content types, though that is not its original intent.
All DUA page-level clicks
This table shows all events from all DUA pages. You can filter it by page, click type, (event target in GA) or action or link click (event label in GA).
Clicks from mass.gov to transactional apps (only)
This table shows just outbound link clicks by page from DUA content to the transactional apps and the ui-cares-act.mass.gov site. You can filter it by click URL (event label in GA). There's also a scorecard and a line chart that summarize this data.
These charts allow the CX team to explore the actions taken on the page types they feel are most likely to drive users towards the desired action (completing applications).

Feedback survey data

This page is an expansion on the Nos per 1,000 scorecard on the Summary page. You can filter by page.
Providing a quick overview of responses to the survey for DUA pages allows the CX team to better prioritize content updates.

Login pages

This report includes descriptive data--pageviews, time on page, exits, etc. -- from UIOnline and unemployment.mass.gov's landing and exit pages. In both cases, a table shows a page-by-page breakout, and a line chart shows a summary.
A more detailed view that includes a broader scope of how users begin and end their journey. This can help the CX better understand if people are entering on pages that aren't as ideal as others, which may lead to premature exits.

User Actions from GA

Transactional app form submits & landing pages that lead to form submissions
This page is focused on UIOnline and unemployment.mass.gov form submissions and what page users landed on that led to those submission. Two scorecards totaling the form submits for each transactional app are provided along with a table breakdown for both by landing page
Returning user actions on UIOnline
This table shows the actions taken by users returning to UIOnline. Keep in mind that a returning user could've done anything in previous sessions; we're not looking at a segment of previous submitters here.
You can sort it by language or by event label (which shows the click text).
DUA contact information clicks on www.mass.gov
Shows contact info clicks by Mass.gov page. You can sort by source/medium to see if some segments have a greater need for contact info clicks.
This page is another lens for helping the CX team understand if users are entering the site appropriately, taking intended actions and how landing on certain pages can produce better journey outcomes than others.

GA Form Submits not in Mass

Lets you look up form submissions by city and state (other than the ones in Massachusetts). You can filter by date, city, state and UI/unemployment.mass.gov task path (event category).
This can help detect fraud or other false submissions (such as developers testing the site, etc).
Last modified 7mo ago