A host of bench and engine tests have historically been used by formulators to assess fuel economy when developing engine oils for gasoline-powered passenger cars and light trucks. Some of these methods assess basic lubricant physical properties such as hydrodynamic, boundary and thin-film friction, and are useful for quickly screening experimental components and formulations. Some methods assess rotational drag of a motored engine and offer insights into the friction of various engine parts. Still other methods directly measure the energy consumption in a test engine running in a research laboratory and thus come the closest to simulating a consumer-operated vehicle. Each test method has inherent limitations and is based on underlying assumptions, producing artifacts that must first be understood and then analyzed for relevance to either industry lubricant specifications or real world fuel economy performance. In this paper the authors compare the response of the various test methods to typical engine oil formulation approaches, and evaluate the correlation between test response and real world fuel economy performance as measured in a specially designed vehicle fuel economy test. The authors show that correlation may be established between research tests and standardized fuel economy tests such as the Sequence VID, but the correlation is not as straightforward when applied to performance in a real world vehicle fuel economy test due to the myriad engine and vehicle types available in today's market. Therefore, engine oil formulation approaches for GF-6 and beyond will need to take into account both performance in standard engine tests and response in various types of consumer vehicles in order to achieve maximum fuel efficiency in real world driving.