Background: Problematic alcohol consumption is associated with a high disease burden for affected individuals and has a detrimental impact on companies and society due to direct and indirect health costs. This protocol describes a study design to evaluate the (cost)-effectiveness of a guided and unguided Internet-based self-help intervention for employees called "GET.ON Clever weniger trinken" (be smart - drink less) compared to a waiting list control group. Methods: In a three-arm randomized controlled trial, 528 German adults who are currently members of the workforce will be recruited by occupational health departments of major health insurance companies. Employees aged 18 and older displaying problematic drinking patterns (>21/14 drinks per week and an AUDIT score∈>∈8/6 for men/women) will be randomly assigned to one of three following study conditions: 1. unguided web-based self-help for problematic drinking, 2. adherence-focused guided self-help, and 3. waiting list control. Self-report data will be collected at baseline (T1), 6 weeks (T2), and 6 months (T3) after randomization. The primary outcome will be the reduction of alcohol standard units during the 7 days prior to T2, using the Timeline Followback method. Cost-effectiveness analyses to determine direct and indirect costs will be conducted from the perspectives of employers and the society. Data will be analyzed on an intention-to-treat basis and per protocol. Discussion: There is a need to identify effective low-threshold solutions to improve ill-health and reduce the negative economic consequences due to problematic alcohol drinking in workforces. If the proposed web-based intervention proves both to be efficacious and cost-effective, it may be a useful tool to increase utilization rates of interventions for problematic drinking in occupational settings. Trial Registration: German Register of Clinical Studies (DRKS): DRKS00006105, date of registration: 2014-07-07.