An important feature of group insurance and retirement benefits is that employers must design them around collective — rather than individual — employee characteristics, needs, and preferences. Given the analytic importance of linking data about the benefits offered by employers to data describing the characteristics of their workforces, and the severe limitations of existing data in this regard, RAND was commissioned to study the design and feasibility of an appropriate survey. This study suggests doing a pilot test to evaluate both design and feasibility issues. Such a workforce survey will require a great deal of discipline on the part of its designers to avoid asking more of employers than they are willing to give. The survey will probably not achieve the response rates from employees that are customary in other household or employer surveys. And such a survey will be relatively expensive, requiring a corps of experienced, skilled interviewers and flexible data-collection procedures that are tailored to each employer.