Speed-Tests: Substitute for, or Complement to, Broadband Maps?

The Federal Communications Commission’s existing broadband availability maps have been heavily criticized as inaccurate, especially for the purpose of distributing billions in subsidy dollars to extend broadband networks to unserved areas. In a rush to distribute the National Telecommunication and Information Administration's (NTIA's) Broadband Equity, Access and Deployment (BEAD) subsidy dollars, a few states have initiated their own mapping efforts and some advocates have proposed alternative mapping means using speed-test data to identify areas that lack adequate broadband. The usefulness of speed-test results to identify unserved or underserved areas has not been evaluated, and there are several reasons to doubt their utility:

• Speed tests reflect what speed tiers consumers buy rather than what is available;
• Tests are convenience samples;
• Testing is often used for troubleshooting, which biases the results downward. 

The results are not encouraging. Speed tests do not appear to be good proxies for actual broadband deployment. In areas where the Ookla data indicate broadband speeds are below the 25/3 Mbps threshold, about 86% of locations have broadband service available above that threshold. At the 100/20 Mbps threshold, about 94% of locations may purchase service meeting or exceeding that threshold. Other peculiarities are found, suggesting that speed-test results are not a suitable substitute for rigorous and careful mapping efforts, such as those the FCC is currently undertaking.