This reminds me of election polls in the UK. Before an election some agency will phone up a thousand or so people and ask them what they plan to do on election

This reminds me of election polls in the UK. Before an election some agency will phone up a thousand or so people and ask them what they plan to do on election day - tick the red box or the blue box. Sometimes they also go door to door or ask people in shopping centres. Despite their methods they do get to miss a lot of folk - those that do not have phones and those that are out working rather than waiting for the pollster to ring their doorbell. Consequently the results are not necessarily representative of true voter intentions. As for the total XP installed count, what about all those machines that get used every day in some workplace for something as ordinary as printing labels but never get to go online? Or those rooms full of PCs that only get occasional use because people bring their own device or do their personal surfing on a phone? Or those PC's in back rooms that generally collect dust? These machines are like the voters the pollsters miss.