IDC and Appcelerator released their Q1 survey of developer sentiment this morning. Below, I have copied the summary chart of the survey, the share of developers who are very interested in developing for each platform.
The report claims “Apple shines, Android slows” but notes “Although technically within standard deviations, these drops stand in contrast to steadily increasing developer interest in Android over the last year and are consistent with an increase in developer frustration with Android.” These data have led to a host of misleading and false headlines this morning: Android Developer Interest Wanes.
If the results are within two standard deviations, it’s impossible to conclude that Apple or Android developer sentiment has changed in any meaningful way with 95% confidence. Simply reading the headlines of these stories would lead an Android developer astray. Interest in iOS seems to fall as much as Android as a percent of the total. Maybe there’s a seasonal effect or maybe there’s a difference in the wording of the question – who knows.
The point is bad data lead to bad decisions. We must interpret data rigorously both at the industry level and within a company. If your company has a data set to analyze, like many of our companies, find an advisor or an employee who is skilled in statistics. You will make better decisions.
When I was at Google, we had many of these experts, Hal Varian leading them. Because of their expertise, our product teams made decisions with additional confidence knowing that our conclusions from the data analysis of advertiser success or publisher performance or customer satisfaction surveys were sound.