Public programs often struggle to reach the people they are meant to help. In this blog post, Christian Schimpf, Vince Hopkins, Priscilla Fisher, and Jeff Dorion explain how behaviourally-informed email messages can successfully encourage enrolment in employment services. Their experiment doubled participation, showing how small design choices can improve access and help address low take-up of public programs.
Few people use public programs when participation is voluntary
Across the world, many public programs struggle to reach the people they’re designed to help. Even when services are free, many people never enrol. This is known as the take-up gap: the difference between those who are eligible for a program and those who actually participate. And while this gap varies across programs and countries, in many cases it is surprisingly large.
The take-up gap is especially large when participation is voluntary. Unlike mandatory programs, voluntary ones require people to consider whether they are eligible, assess the potential benefits, and take the time and initiative to apply. When those steps are confusing or burdensome, participation drops.
Low take-up poses a serious challenge for governments. When public programs go underused, the resources invested in them may be wasted, and the social and economic benefits they are meant to deliver can go unrealized.
The question for policymakers is: when programs are available, what prevents people from participating? And more importantly, what can policymakers do to close that gap?
Why so few people use employment services
Public employment services are designed to help people find work, upgrade their skills, or connect to training opportunities. Yet, like many other public programs, they often struggle to reach the people they’re meant to serve. In British Columbia, Canada, earlier work has shown that take-up of these services is low, with participation as low as one percent among recently unemployed individuals.
To understand why, we partnered with the Government of British Columbia, which oversees employment programs across the province. These programs include an online job bank and a network of employment offices run by local service providers. Drawing on academic research, interviews with jobseekers and staff, and a national online survey, we found that many eligible people face barriers both getting ‘to the door’ of these services and getting ‘through the door’ once they try to enrol.
By barriers getting to the door, we mean the challenges that prevent people from applying for these services at all, such as not knowing what services are available or facing financial pressures that make it hard to plan ahead. Getting through the door refers to what happens after people become interested. For some, the online enrolment process is challenging, which can be enough to discourage them from participating altogether.
This distinction helps clarify where and why participation breaks down. Some people never make initial contact, while others start the process but drop off before completing it. Recognizing this difference can help policymakers target their efforts more effectively, whether by improving outreach efforts, simplifying enrolment procedures, or both.
These insights raised our next question: could behaviourally-informed design changes make it easier for people to enrol in employment services?
Designing and testing behaviourally informed interventions
To answer that question, we worked with the Government of British Columbia to develop a set of low-cost, behaviourally informed interventions that addressed different barriers people face when deciding whether to enrol.
All of the interventions provided the same email: a short, clear overview of what employment services offer and how they could help. Beyond that, we introduced two behavioural variations:
1. Call to action: People in the ‘standard choice’ prompt saw a single button to click on. In contrast, recipients of the ‘active choice’ prompt were encouraged to make a decision about enrolling right away through highlighting the choice they have with two buttons, a ‘Yes’ and a ‘No’ button.
2. Hand-off: People in the ‘cool hand-off’ group led those clicking through the email to the standard online enrolment form. In contrast, people in the ‘warm hand-off’ group were given the option to connect directly with their nearest employment office for personal assistance through an expression of interest form.
Together this allowed for four experimental groups to be compared with a control group, which received no email at all (see Figure 1).
We evaluated our interventions through a randomized controlled trial with 9,877 unemployed jobseekers between 11 September and 12 October 2023. Everyone in the study had recently applied for Employment Insurance, Canada’s unemployment income support program. Each person was randomly assigned to one of the five groups.
We then measured two important outcomes: did the participant click-through and submit an expression of interest (to the door)? And did they enrol in the employment services (WorkBC) within a 14-day period (through the door)?
Figure 1: The evaluation design
Source: Schimpf et al (2025)
Did behaviourally informed interventions increase take-up of employment services?
The short answer is “yes, but…”. Our best performing interventions – those with the cool-handoff – more than doubled the number of people enrolling compared to the control group, who received no email at all (see Figure 2). That’s an encouraging result.
Figure 2: Evaluation results
Source: Schimpf et al (2025)
But not every intervention worked as expected. The “warm hand-off,” which invited people to connect directly with their nearest employment office, did not increase enrollment compared to the control group. Why might that be?
Looking more closely at the data, a clear pattern emerged. Many people were interested in the program – roughly one in ten clicked on the link in the email (‘to the door’). But only a small share went on to enrol (‘through the door’). In the end, only about one in a hundred of the people we emailed actually completed registration for employment services.
This drop-off was consistent across all groups in the trial. Notably, however, the warm hand-off interventions performed slightly worse: their conversion rates were about five percentage points lower than the cool hand-off versions. In short, offering direct contact with an employment office did not make it easier to enrol; it actually seemed to create extra friction.
Converting interest into enrolment remains a key challenge. Even seemingly small barriers can determine whether interest translates into enrolment, reminding us that design details matter, even in programs that are free to access.
Implications for policy implementation
Low take-up continues to be a major challenge for policymakers. Our partnership with the Government of British Columbia showed that there’s rarely a single reason why eligible people do not participate – and therefore no single solution. Participation gaps usually arise from a mix of factors, such informational, financial, and procedural barriers.
Behaviourally-informed adjustments to the enrolment process can help, but their impact has limits. Not every intervention will succeed, which is why grounding policy changes in evidence and testing them carefully is so important.
Our findings highlight the value of distinguishing between getting people to the door and getting them through the door. This lens helps program designers and evaluators pinpoint where participation breaks down. In our case, lack of interest wasn’t the problem; the challenge was converting that interest into enrolment.
For policymakers, this distinction offers a practical guide for allocating effort and resources. When people are not getting to the door, governments should invest in marketing and communications to improve clarity and raise awareness. When people cannot get through the door, the focus should shift to improving the application process, through simpler forms, fewer steps, and more intuitive design.
Read more about the research here.
Christian H. Schimpf is a Postdoctoral Research Fellow in Political Science at the University of British Columbia and member of The Behavioural Public Policy Lab
Vince Hopkins is an Assistant Professor in Political Science at the University of British Columbia and PI of The Behavioural Public Policy Lab
Priscilla B. Fisher is Research Manager of the Stone Centre on Wealth and Income Inequality and the Centre for Innovative Data in Economics Research at the Vancouver School of Economics, University of British Columbia
Jeff Dorion is the Chief Data Officer in the BC Public Service Agency at the Government of British Columbia