Over the last few years, data collection agencies have gained experience with asking for linkage consent in face-to-face and telephone surveys; however, the same is not true for web surveys. Asking for linkage consent in a web survey raises important methodological questions, namely, what is the optimal way of asking for linkage consent that maximizes consent rates and minimizes consent bias? We addressed this question using a national probability sample of people living in Germany. Web survey respondents were asked for consent to link their survey data to official employment records. Respondents were randomly assigned to receive different wordings of the consent question. We found that when the consent request is phrased in such a way that emphasizes a particular benefit of linkage (reduced questionnaire and time savings) respondents provide consent at a higher rate compared to neutral phrasing.