An efficient sampling design for forest inventory is important to collect the basic measurements required to estimate stand volume and describe forest structure. For operational forest inventory, efficient sampling designs often are constrained not only by standard error objectives, but also by cost constraints. So, sampling designs must consider a balance between objectives and constraints. Big basal area factor (BAF) sampling incorporates subsampling into horizontal point sampling by using two angle gauges: one to select count trees (small BAF) and one to select measure trees (big BAF). Big BAF sampling focuses inventory effort on counts, which tend to be more variable than volume:basal area (VBAR) measures, and greatly improves sample efficiency. However, big BAF sampling is not widely employed because forest inventory specialists do not fully understand the relationships among sampling error, sample size and field costs. In this study, we show the impact of count BAF and measure BAF on sampling error and sample size requirements, and develop methods to optimize choice of count and measure BAF when inventory expenditures are constrained. Our results show that when the variability in VBAR is low, count BAF limits sample design options, and as VBAR variability increases, measure BAF becomes the limiting design factor. We also found that efficient sample design is more critical when inventory resources are limited than when inventory resources are more abundant.