In the fall of 2018, the Digital Transformation Office (DTO) worked with Employment and Social Development Canada (ESDC) teams on two optimization projects: Student loans and grants, and Employment insurance for maternity and parental benefits. Both projects involve top tasks that millions of Canadians seek to complete on regularly. Both projects produced evidence of simple content design principles that can greatly improve user success.

1. Remove layers, help people get answers faster

While the Government of Canada funds student loans, provinces and territories manage them. Students use the provincial sites to apply and reapply. They only need to interact with the federal government when it is time to repay a loan. had many pages describing funding programs. The project team stripped these pages out and created clear paths to the pertinent content on provincial sites and the National Student Loans Service Centre site. Removing pages that didn’t need to be there, meant the path to answers was shorter. Participants took an average of 3 fewer clicks to get to the destination content on our redesigned prototype.

Recreating information that other organizations manage can bloat your content. This makes it harder for users to find what they are looking for. Creating clear “off-ramps” where possible and reducing clutter and noise improves user success.

2. Group content by task, not by which team is responsible for maintaining it

One of the challenges of the Student loans project was the way content for grants and loans was divided. Because different program teams manage this content, the information was on separate pages in the site. This didn’t make sense for students though. They apply for both loans and grants in a single application form. In the redesigned prototype, the project team combined grants and loans in a single topic page. This made the path and process much clearer.

Getting a clearer understanding of the mental model your users have can help you group content appropriately. Challenge internal departmental silos. While these help divide the work of managing specific programs, they often complicate the user journey. Service design requires us to talk to colleagues beyond our own teams and to walk in our users’ shoes so we can design elegant user journeys that simply make sense.

Before and after web page for student loans.
Grouping grants and loans together made the path to the task much clearer. Using a steps pattern on the topic page helped narrow options quickly. The result: less cognitive load and greater user success.
Detailed description

There are images of the original Student loans page and the Student grants from the baseline round of testing. A green arrow points from them to the prototype version of the same content. The prototype version is called Student aid. At the top of the page is the subheading: Student grants and loans. Below this are 3 linked options:

  1. Apply for student grants and loans
  2. Manage your loan at the NSLSC
  3. Repay your student loan

3. Use evidence to design language

The team wanted to understand how students were searching for information. They looked at Reddit posts, search data from within and Google search queries and trends. While one of the main topic pages was called “Student financial assistance,” the research showed that students search for “Student aid.” So we changed the name of the page. We also learned that the term “lump-sum payment” was better understood and more accurate than “one-time payment.” So we updated that terminology in the prototype.

We often forget that people beyond our own teams don’t work with our lingo on a daily basis. Testing assumptions around terminology, and doing our research ensures that our audience will better understand our message.

4. Don’t make people do math

In the first round of our EI maternity project, only 5% of participants were able to correctly calculate the maximum amount of paid leave a mother could take after giving birth. The content had all the numbers you needed, but people struggled with how to put them together. A note pad, pencil and arithmetic skills were required to figure out the basic question every parent is asking: “How long can I stay home with my baby?”

The project team attacked this problem on two fronts. First, we made an estimator using an interactive wizard pattern. This let people enter some basic data and get an estimate of the number of weeks and the amount of money they could expect. For those who prefer to see an overview of the options, we took the numbers out of paragraphs and created a simple table. That let people see at a glance the weeks, benefit rates and related dollar amounts for each benefit option. Below the table, we included 2 examples that illustrated how to do the addition. Showing people how to do the math and giving them a tool to help them do it increased task success for this question by 74 percentage points. (psst…5% + 74 points = 79% success).

People don’t math

Before page for the number of benefit weeks.
People didn't even consider the benefits being used in combination. They focused on the higher number, thinking it was the total.
After page for the number of benefit weeks.
Showing a formula people can use and relate to helps them do the math.
Detailed description

The image first shows a section of the original page content with red highlighting around a statement about a maximum of 15 weeks of EI maternity benefits. There is red highlighting around a second statement about a maximum of 61 weeks.

Below this is a second image from the prototype version of the same content. There is green highlighting around a simple math equation. The equation shows 15 weeks maternity plus 61 weeks of extended parental equals 76 weeks total for Janelle.

5. Use numbers sparingly

People find numbers on websites hard to digest. On this project we found many content pages peppered with numbers. Numbers of weeks, benefit rates, qualifying periods, dollar amounts, and numbered footnotes. On one important answer page we counted 30 different numbers in just 3 paragraphs of text.

Simplifying content doesn’t only apply to the words on a page. If people are looking for the answer to a “how much” question, they gravitate to the first or the biggest number they see. If content is riddled with numbers, task success is very inconsistent. Too many numbers confuse people. Be very careful how you use them. Clear, simple statements of facts with minimal numbers make for more consistent success rates.

Too many numbers equals overload

Before and after benefits page.
Clear, simple statements of facts are far more successful
Detailed description

On the left is a screenshot of three dense paragraphs of text from the original content. There are red boxes drawn around each of 30 numbers included in the text. A green arrow points to two screenshots on the right of the redesigned content in the prototype. Under the heading "Special circumstances" is a small paragraph of text in an expand/collapse field. A green box highlights a simple sentence with a link to EI sickness benefits. Below this is a second heading: Eligibility. This has a similar expand/collapse field with a similar small paragraph of text. Again, a green box indicates the same information with a link to EI sickness benefits. Below everything is a green box indicating that these changes improved results 58 percentage points.

6. Put answers where people look for them

This is advice we repeat in almost every optimization project. It seems so simple, but it’s not always obvious. Our own deep understanding of our subject matter can get in the way. What’s obvious to us doesn’t always match how our users think and navigate. In the EI parental benefits project, one of our tasks asked whether a mother has to use her maternity leave if she becomes ill and must stop working before her baby is born. The answer is no, she can apply for EI sickness benefits. In our baseline testing this task had only a 22% success rate. We saw what we call “pogo-sticking.” People bounced from one page to another, trying to figure out where to find the answer. None of the page titles provided a clear scent of information for this scenario.

In the prototype, the project team introduced a page called “Special circumstances” and moved the answer to this new page. The majority of participants in the validation round went to this page to solve the task. It fit their mental model. We also looked at our baseline data to understand where else people might look. We saw that the eligibility page had been the most popular. So we also put the answer there in our prototype. With these two changes, success for this task soared to 80% in our validation round.

Final word

Optimization projects are about connecting the lived experience of our users to how we design our content. Having a deeper understanding of how people navigate and consume our content gives us important insights. When applied to content design, these insights help us provide better services to Canadians.

We want to hear from you

Let us know what you think. Email us at or tweet using the hashtag #Canadadotca.

Learn more