Food Cost Variances in Regional Chains

Many of my regional restaurant clients show significant differences in the food cost % from unit to unit. It is always easy to find anecdotal excuses for these variances. Examples include: customers at each location have different dining habits, higher sales volume hides some mistakes which surface at lower volume locations, and the higher volume locations have better personnel. Which units should be used as a benchmark? Should the lower volume units be par? Or are these units below the standard? For purposes of growing a chain, I would use the higher food cost % in projections. If a chain always look at the best case scenario in five year plans, they will never hit their target figures due to rosy assumptions.