> Twitter said its data-retention policies made it unable to reconcile the figures for periods before last year’s fourth quarter.

Is Twitter really claiming that they do not store even aggregated data on number of users for more than a year? Or does the data prior to that not have any distinction between Twitter and third-party users that could be used to correct their error? I would be amazed if a social media platform of that size had such weak retention policies, even for their own sake. Maybe I am too used to the paradigm of Google/FB where data is gold stored in Fort Knox.