If there is one thing uniting the diverse missions of people working in the non-profit and public sectors, it’s that everyone wants to make a difference. Whether that’s the Centre for Homelessness Impact, local authorities, the charities we work with, or volunteers on the ground, the shared goal is to improve lives. To do so effectively, using evidence and data is indispensable. This is because while efforts to help people are almost always well-intentioned and often based on good intuition and previous experience, many of our good intentions might not improve outcomes or be as effective as we hoped. In fact, a reasonable rule of thumb is that 80% of programmes don’t work (see discussion on this here and here). This can be difficult to accept when so much effort has gone into trying to make an intervention impactful. But once we acknowledge that a big part of ‘What Works’ is finding what doesn’t work, we can see the big opportunities this provides for learning and improvement. Importantly, this acknowledgment allows us to shift scarce resources away from less effective interventions, and retarget them to areas where they do have an impact. 

Testing an email support intervention

Recently, we trialled an intervention that did not have the effect we hoped for. We set out to address a pressing issue: burnout in the homelessness workforce. Frontline staff working with people experiencing homelessness have lower wellbeing than other workers, as well as facing higher burnout and low retention rates. This is due to a variety of factors, including exposure to high-stress situations and worsening working conditions. 

Working with St Martin-in-the-Fields Charity and our evaluation partners at Exogeneity and the London School of Economics, we set out to trial an intervention designed to improve wellbeing and reduce burnout for these workers. Based on promising evidence from the United States, we sent social support messages directly to frontline workers via email. In these messages, frontline workers shared stories about how they felt their work had made a difference, and reflections on their experiences. By sharing these stories, the goal was to increase feelings of belonging and social connection, and ultimately reduce burnout and increase wellbeing. 

We ran a randomised controlled trial to robustly test the impact of this support. Our results did not, however, find strong evidence of a positive impact. Two potential reasons emerged for why the emails didn’t make the impact we hoped for. 

1. The intervention may not have been appropriate for frontline workers in the UK – the intervention was adapted from a promising example in the US which was targeted at 911 operators. When tried with UK homelessness staff, participants pointed out that their workload, particularly the high volume of emails that frontline workers in the homelessness sector already received, may have prevented them from engaging with the intervention in meaningful ways. 
2. The intervention may have been too light-touch to meaningfully move the dial on wellbeing. While participants were positive about the emails and appreciated that it gave them a sense of community, it was not perceived as enough to counter the difficult environment they are working in and meaningfully improve their wellbeing. 

While we always hope to see positive impacts from these interventions, and it’s disappointing when they don’t have the impact we expected, there are huge opportunities to learn and improve our approach going forward. 

The humility of asking ‘what works’  

In difficult global and local circumstances where time and money are tight, it’s important that the scarce resources we do have are directed where they can make the biggest difference. By learning that this intervention didn’t have a significant effect on the wellbeing of frontline workers, we can ensure it doesn’t continue over time and allows those potential savings to be redirected towards other interventions that might make a meaningful difference to these outcomes. 

These findings also give us an opportunity to understand where interventions aren’t quite appropriate, or need tweaking to have the impact we’re after. We had promising evidence to suggest this intervention would work – 911 dispatchers receiving a similar intervention had significant reductions in burnout and reduced resignations by half. But the workload and volume of emails frontline homelessness workers receive was probably one of the main barriers to this intervention having the desired impact in our context. Workers reported finding it difficult to meaningfully engage with the support due to their high workload. 

Another key takeaway from this trial is the iterative nature of understanding. While we begin with contextual awareness, the process of testing itself often uncovers nuances and assumptions we hadn't fully appreciated, guiding us toward a more effective approach. The light-touch approach was important to test, since if it was successful it would have been a low-cost, scalable intervention that resource-poor and resource-rich charities alike could have rolled out. But the lack of positive effects underscores the fact that support for frontline workers must account for the wider context that they operate in, and the magnitude of the challenges they face. 

Light-touch approaches may be insufficient to meaningfully improve their wellbeing, when they are dealing with high workloads, a precarious sector, and regular exposure to high-stress situations. An approach that looks at the overarching structures these workers operate in likely holds more potential for improving their wellbeing and reducing burnout.

Making a meaningful difference

These lessons all help us move towards a better understanding of how we can move towards our ultimate goal of meaningfully improving the wellbeing of frontline workers. What unites everyone involved in this project is this desire to improve the lives of staff who make such a huge difference in the lives of so many people, and make up the backbone of the homelessness sector. 

Making these meaningful changes is a journey, and the findings from this trial help us take a vital step forward. We need to actively commit to an evidence-based approach to further progress down this road. We must not just create new evidence, but decisively use this evidence - whether the findings are positive or otherwise - to fuel our progress. We must reframe null or negative findings as valuable lessons rather than failures. 

It’s time to take on this approach, rigorously test promising avenues, and redistribute resources to where the evidence shows us they can have the biggest impact. By doing so, we will make the improvements that enable us to make a more meaningful difference in the lives of those we are dedicated to supporting. 

•  Luke Arundel is Evidence and Data Lead at the Centre for Homelessness Impact