Over the last 20 years, diagnostic fracture injection tests (DFITs) have evolved into commonly used techniques that can provide valuable information about the reservoir, as well as hydraulic-fracture-treatment parameters. Thousands of tests are pumped every year in both conventional and unconventional reservoirs. Unfortunately, many tests that are pumped provide poor or no results because of either problematic data acquisition or incorrect analysis of the acquired data. This paper discusses common issues and mistakes made while acquiring DFIT data. Guidelines on how to avoid these errors and secure the best possible data are provided, including data resolution, pump rates, test duration, and fluid selection. Strategies are provided to estimate the time required to reach fracture closure and establish stable reservoir transients for analysis. The last part of the paper addresses potential (and commonly observed) problems in the analysis of the DFIT. These issues can be magnified in tight-gas and shale reservoirs because of the long data-acquisition times and the subtle pressure transients that can occur. Specific issues that are discussed include poor instantaneous-shut-in-pressure data from perforation restriction, loss of hydrostatic head, gas entry and the resulting phase segregation, the use of gelled fluids, and errors in after-closure analysis.