Statisticians are under pressure to innovate, partly due to shrinking budgets and the call to do more with less, but also due to technological advances and emergence of new actors promising to produce more accurate and timely statistics with what has come to be known as “big data”. This raises the question, how do new forms of data and methods become legitimate and official? We approach this question by conceiving of official statistics as part of a transnational field in which different factions of actors compete and struggle over the authority to innovate the data and methods that are legitimated to produce official statistics. We consider these struggles as a politics of method that is not reducible to a competition between ideas and words. They are also material insofar as they feature competing digital devices mobilised to demonstrate the validity of new data and methods. Through two empirical examples, we identify the strategy of reassembling methods to capture how statisticians tame and contain innovations based on big data, especially those introduced by data scientists, by integrating and simultaneously subordinating them to existing methods. By doing so, we suggest that reassembling is an innovation strategy that secures the relative position of national and international statisticians within the transnational field of statistics.