Assay design is an important variable that influences the outcome of an inhibitor screen. Here, we have investigated the hypothesis that protein tyrosine phosphatase inhibitors with improved biological activity could be identified from a screen by using a biologically relevant peptide substrate, rather than traditional phosphotyrosine mimetic substrates. A 2000-member library of drugs and drug-like compounds was screened for inhibitors of lymphoid tyrosine phosphatase (LYP) by using both a peptide substrate (Ac-ARLIEDNE-pCAP-TAREG-NH2, peptide 1) and a small-molecule phosphotyrosine mimetic substrate (difluoromethyl umbelliferyl phosphate, DiFMUP). The results demonstrate that compounds that inhibited enzyme activity on the peptide substrate had greater biological activity than compounds that only inhibited enzyme activity on DiFMUP. Finally, epigallocatechin-3,5-digallate was identified as the most potent inhibitor of lymphoid tyrosine phosphatase activity to date, with an IC50 of 50 nM and significant activity in T-cells. Molecular docking simulations provided a first model for binding of this potent inhibitor to LYP; this will constitute the platform for ongoing lead optimization efforts.