What WWII Bomber Design Can Teach You About Interpreting Data
Heavy bombers were critical to Allied strategy of destroying the Axis industrial base, but in an era without smart bombs and guided missiles, the strategy required WWII bombers to deliver tons of ordnance to get the job done - and many of those bombers weren’t returning from missions.
Their solution was to increase the armor of each aircraft; more protection meant a better chance of completing its mission and returning to base. But armor came with a performance penalty- increased weight. Adding more armor meant fewer bombs could be carried to compensate.
Abraham Wald, a mathematician, tackled the problem by graphing the location of bullet holes in returning bombers. Conventional wisdom dictated that those areas required additional armor.
However, Wald interpreted this data differently. He concluded that the planes who returned to base were notable for where they didn’t have bullet holes. A plane could obviously take damage on the wings and fuselage (where the returning planes had damage), but adding armor to the cockpit area would increase the number of planes safely returning to base.
Nate Bolt, a research manager at Facebook, uses this story as reminder to constantly evaluate the copious amount of data generated by users. Sometimes the holes in the data are as important as the data you receive.