These include: adapting program logic models, mixing evaluation methods that have different intents, shifting how we think about the utility of evaluation, and rightsizing evaluations. Adapt, Do Not ...