In an effort to rein in rising health care costs, health plans are using physician cost profiles as the basis for tiered networks that give patients incentives to visit low-cost physicians. Because physician cost profiles are often statistically unreliable some experts have argued that physician groups should be profiled instead. Using Massachusetts data, we evaluate the two options empirically. Although we find that physician-group profiles are statistically more reliable, the group profile is not a good predictor of individual physician performance within the group. Better methods for creating provider cost profiles are needed.