Annual vaccination is the most effective way to prevent seasonal influenza. However, globally, the recommendations vary from country to country, ranging from universal recommendations, risk‐group‐specific recommendations, to no recommendation at all. Due to high diversity both in recommendation practice and country‐specific preconditions, it is difficult to determine the effect of different recommendations on vaccine uptake. This incentivised laboratory experiment (N = 288) tests the behavioural consequences of different recommendations in a repeated interactive vaccination game. The participants are part of heterogeneous groups, comprised of low‐ and high‐risk type of players. They receive either a universal, risk‐group‐specific or no recommendation prior to their vaccination decisions. Results show that individuals are sensitive to the recommendations. In detail, a risk‐group‐specific recommendation increases vaccine uptake of high‐risk types. However, at the same time, it decreases vaccine uptake of low‐risk types. The results imply that when the proportion of low‐risk types in a population is considerably larger than the high‐risk group, a risk‐group‐specific (vs. universal) recommendation comes at the cost of decreased social benefit of vaccination due to the overall lower vaccine uptake. Policy decision‐making should therefore complement epidemiological considerations with potential positive and negative behavioural consequences of vaccination recommendations.