For many who search for jobs in occupations with poor prospects, the most promising option may be to consider a career switch. But without the ‘relevant experience’ for the jobs they are considering applying for, job seekers are likely to be out of their comfort zone. The challenge of supporting job seekers who have to broaden their search is particularly relevant as economies try to adjust to the digital and green transitions (Van Herck et al. 2026, Causa and Soldani 2025, Sahin et al. 2014). Our research (de Koning et al. 2026) provides a ‘proof of concept’ for a tool that addresses these challenges by leveraging the experience of recent job seekers who have successfully switched occupation. The tool consists of a digital platform that allows job seekers to connect to former job seekers who volunteered to serve as a ‘buddy’. The evidence suggests that such tool may be a promising additional resource for job seekers, complementary to existing public support schemes that are currently under fiscal strain in many countries.

Interventions that have been tried

Traditional tools such as training programmes and job-search requirements have shown mixed results. While training programmes for the long-term unemployed can be effective (Eppel et al. 2024), their efficacy strongly depends on programme design and the costs may be high (Wood et al. 2025). The importance of design is also true for job-search interventions. Vethaak and van der Klaauw (2022) show that mandatory broader search requirements in the Netherlands reduced job-finding outcomes, suggesting that pushing job seekers to apply more widely can backfire when it ignores their situation and incentives. By contrast, Kircher et al. (2016) and Bachli et al. (2025) study personalised occupational recommendations and find improved outcomes for job seekers. 

Effective job search requires knowledge and experience of how best to search. In this process, social connections can be helpful (Lalive et al. 2023), but research has shown that unemployed job seekers are not well connected to the employed (Chetty et al. 2022). Matching them with buddies could enhance social support through enlarging their network and helping them navigate the job search process (Caliendo et al. 2015). They could also help deal with psychological barriers associated with seeking new employment or act as enforcers of social norms around job search (Kondo and Shoji 2019). 

The idea: Learning from someone who’s been there 

Our idea is simple but novel: pairing unemployed individuals with ‘buddies’. The premise is intuitive. Job seekers may benefit not only from formal advice from public employment services but also from practical, experience-based guidance, encouragement, and social support from peers who have recently navigated similar challenges. Our intervention uses an online platform to connect unemployed job seekers with these volunteer ‘buddies’. These buddies are not professional counsellors; instead, they are individuals who had recently exited unemployment and found jobs in new occupations. This distinction is important. Information can tell job seekers that alternative occupations exist. But information alone may not be sufficient when job seekers are concerned about what a switch would mean for their identity, whether they would fit in a new occupation, or how to take the first practical steps. A buddy can potentially help on all three margins. Such peers can make alternatives feel more attainable, help interpret what a transition looks like in practice, and provide encouragement from someone with whom job seekers can more readily identify. 

The programme specifically targets unemployed job seekers in occupations with poor employment prospects who may need to consider switching careers. Buddies, in turn, are selected because they successfully made such transitions. This matching is crucial: it ensures that advice is relevant and grounded in real-world experience. The key question is whether such a relatively light-touch, and therefore scalable, intervention can meaningfully improve employment outcomes. 

The experiment: Testing the impact 

To answer this question and evaluate the promise of such tool, we designed and evaluated a small-scale randomised controlled trial in collaboration with the Dutch Public Employment Service. A group of unemployed individuals were randomly assigned access to the online buddy platform, while a control group continued with standard services. The platform allowed job seekers to connect with a buddy and to exchange messages and receive guidance on job-search strategies, skill transitions, and navigating labour market uncertainty. The platform also included short videos of former job seekers describing their own transitions. 

This design allows for a clean comparison between those who had access to peer support and those who did not. In total, 713 job seekers participated in the experiment, and our main administrative-data analysis focuses on 524 individuals who were unemployed and not working at all just before treatment. Not all job seekers took up the treatment: 53% created an account and 19% ultimately found a buddy. These percentages reflect the real-world setting where large-scale rollout might face the same constraint that not all job seekers will be open and willing to engage with the platform. Matching with a buddy is strongly selective (for example, it may be particularly those job seekers that have low re-employment beliefs that engage with the platform), and thus all results below focus on the so-called intention-to-treat effect: does getting access to the platform change outcomes? The study tracks participants for 18 months, enabling us to assess not just immediate effects but also longer-term outcomes. 

Main findings: Meaningful gains from access to the platform 

Despite this incomplete take-up, the effects are sizable. Figure 1 shows the basic pattern, based on administrative data. There is little movement in the first year, but the treatment and control groups begin to diverge clearly in months 13 to 18. Over that period, treated job seekers were 6 percentage points more likely to be employed than the control group, an increase of about 11%. Monthly earnings were €226 higher, about 16% above the control-group level. Over the full 18-month follow-up period, cumulative earnings rose by about €2,185. 

Figure 1 Treatment impacts (ITT) on employment and earnings

These effects are notable for several reasons. First, they are achieved through a relatively light touch intervention. Second, they emerge and persist over a fairly long time horizon, consistent with the idea that occupational transitions take time. Job seekers may need time to revise their expectations, broaden their search, build confidence, apply to unfamiliar jobs, and move into employment that is more durable than a rapid but poor match. The strongest results are for those with longer unemployment duration at the start. Figure 2 shows that for job seekers who had already been unemployed for more than 17 weeks when the intervention began, the gains show earlier and are larger. Employment rose by 9 percentage points. Their cumulative earnings gains exceeded €5,000 over the study period. This is an especially important finding for policy because long-term unemployment is where public employment services often struggle most. These results suggest that peer support may be particularly valuable for individuals with weaker labour market attachment. 

Figure 2 Treatment impacts (ITT) on employment and earnings: only individuals with more than 17 weeks of UI duration at the time of intervention 

Occupational switching and other behavioural responses 

The data do not suggest that the buddy platform simply made people search harder. We find no clear increase in the number of applications or in self-reported motivation to search. This points away from a simple monitoring or pressure mechanism. Instead, the evidence suggests that treated job seekers kept searching broadly, while their counterparts in the control group narrowed their search over time. That interpretation also fits the idea that peers may matter less for increasing raw effort than for shaping search strategies, reducing uncertainty, and reinforcing constructive social norms around job search (Kondo and Shoji 2019). 

The initial purpose of the intervention was to assist job seekers in making an occupational transition. Among job finders who answered the follow-up survey, treated individuals were more likely to end up in occupations different from the one they originally targeted. Figure 3 shows that at the detailed four-digit occupation level, 86% of treated job finders entered a different occupation than their initial target occupation, compared with 65% in the control group. The estimates are noisy because this analysis relies on a smaller survey sample, but the pattern is clear and fits the intervention’s purpose. 

Figure 3 Percentage of job finders who entered a different occupation than their search occupation

 Policy conclusions 

At a time when labour markets are undergoing rapid change and many job seekers face uncertain career paths, we highlight the power of a simple idea: people can learn from those who have recently succeeded in finding employment in a new occupation. 

The evidence shows that this can meaningfully improve employment and earnings outcomes, especially for those who need help most. It does not rely on intensive and expensive training, mandated broader search, or financial incentives. Instead, it leverages social learning and behavioural channels that are often underutilised in labour market policy. Since public employment services have access to a continuous flow of job seekers who find employment, implementation of buddy support at larger scale seems feasible. Connecting those workers to current job seekers is not a substitute for caseworkers, training, or income support, but it may be a powerful complement. 

References

Bachli, M, R Lalive, and M Pellizzari (2025), “Occupational recommendations and job finding”, VoxEU.org, 27 March. 

Caliendo, M, D A Cobb-Clark, and A Uhlendorff (2015), “Locus of control and job search Strategies”, Review of Economics and Statistics 97(1): 88-103. 

Causa, O, and E Soldani (2025), “When upskilling is good but not enough: Understanding labour shortages through a job-quality lens”, VoxEU.org, 27 May. 

Chetty, R, et al. (2022), “Social capital I: Measurement and associations with economic Mobility”, Nature 608(7921): 108-121. 

Eppel, R, U Huemer, H Mahringer, and L Schmoigl (2024), “Active labour market policies: What works for the long-term unemployed?”, The BE Journal of Economic Analysis and Policy 24(1): 141-185. 

Kircher, P, P Muller, and M Belot (2016), “How low-cost labour market information benefits job seekers”, VoxEU.org, 10 March. 

Kondo, A, and M Shoji (2019), “Peer effects in employment status: Evidence from housing lotteries”, Journal of Urban Economics 113: 103195. 

De Koning, B K, P Muller, M Belot, Y Engels, D Fouarge, M Keer, P Kircher, and S Phlippen (2026), “Online Buddies for Job Seekers: A Field Experiment”, NBER Working Paper 34912. 

Lalive, R, D Oesch, and M Pellizzari (2023), “How personal relationships affect employment outcomes: On the role of social networks and family obligations”, in Withstanding Vulnerability Throughout Adult Life: Dynamics of Stressors, Resources, and Reserves, Springer Nature Singapore.

Sahin, A, J Song, G Topa, and G L Violante (2014), “Mismatch unemployment”, American Economic Review 104(11): 3529-3564. 

Van Herck, K, A Kiss, and A Turrini (2026), “The rise and fall of EU labour shortages: Recent developments and some forward-looking considerations”, VoxEU.org, 26 March. 

Vethaak, H, and B van der Klaauw (2022), “Mandatory broader job searches for unemployed workers”, VoxEU.org, 16 December. 

Wood, J, K Neels, and S Vujic (2025), “Which training leads to employment? The effectiveness of varying types of training programmes for unemployed jobseekers in Flanders”, Journal of Social Policy 54(2): 651-672.