This paper uses the Longitudinal Employer Household Data (LEHD), matched to administrative data from Workforce Investment Act (WIA) programs in two states to examine the value of firm characteristics as conditioning variables and as outcomes in evaluations of active labor market programs using observational data drawn from administrative records. The paper also acts as a replication of sorts (with a different set of states) of recent work by Heinrich et al. and provides an opportunity to push the "standard" approach in the US literature in several substantive and econometric directions. The primary, and perhaps surprising, finding is that firm characteristics provide (essentially) no additional information relevant to solving the selection problem not already present in the data from Unemployment Insurance (UI) earnings records. They also not provide much insight as dependent variables. More broadly, the estimates prove remarkably robust to the application of the alternative matching and weighting estimators currently occasioning much debate in the applied econometrics literature.