OBJECTIVES: To improve cost-effective patient recruitment via social media for online medical surveys. METHODS: Recruiting patients from social media channels into online medical questionnaires is increasingly common. For every patient completing successfully, many others have failed to engage with the advert or finish the questionnaire. For recruitment to be both successful and cost effective the ratio of those completing to those lost during the process must be maximised. In a social media recruitment campaign for a pain relief gel study we used analytics to record throughput of potential candidates at every stage of the process. At bottlenecks we put a series of A/B tests in place and continuously measured the impact of changes to wording, images and calls to action to throughput at those points. RESULTS: We identified the key points in the patients’ journey where significant losses occurred: advert click through, landing page click through and questionnaire completion. By using A/B testing at these points we were able to improve the landing page conversion rate from 40% to 55% and the completion rate from 45% to 75% over a 7 week period. CONCLUSIONS: Cost-effective recruitment to medical surveys via social media requires ongoing monitoring and adjustment of adverts, landing pages and questionnaires (i.e. continuous A/B testing). These changes cannot be anticipated in advance and require experimentation. Ethical review committees should be aware of this requirement and the need to make small adjustments to adverts etc., without an additional consultation process. Alternatively, flexibility may be built into adverts and surveys submitted to Ethical review committees for approval in the form of options for images, headings and text that may be combined as necessary.