Database attractiveness is independent of data quality.
Global forest products databases have different levels of quality.
Since many data users are not data professionals or statisticians, the quality of a database is not the decisive factor.
A combination of technical, political and developmental factors can explain the increasing discrepancy between the African export and Chinese import forest products trade data.
Global forest statistics providers may be motivated to put more effort into improving database attractiveness and related incentives rather than quality.
What drives discrepancies and inconsistencies in global forest statistics? The use of global statistics has influenced academic research and sectoral policies of forest ecosystems since the first global forest assessment was conducted in 1948 or even earlier. Very little work has been done to provide a comprehensive analysis of the governance structure and the quality of predominant international forest databases. Furthermore, very little is known about the attractiveness and/or repulsiveness of global forest statistics platforms to scholars, policy-makers and other users. To reduce knowledge gap, this article examines the governance structure and strategies of three major databases which provide data on global forest products trade including timber export/import flows data, namely FAOSTAT, the United Nations Comtrade, and Chatham House's Resource Trade Earth. This paper uses conceptual and theoretical frameworks of data governance and nudge theories are used to study the production, quality, attractiveness and repulsiveness of global forest statistics and the related platforms through research on a qualitative and quantitative methodological approach. The main findings show that among the above three data platforms, only Comtrade received first-hand data directly from UN producing member states' offices, while the other organisations depend on Comtrade, transform second-hand data. More importantly, the article reveals that the levels of quality and attractiveness of the forest databases in our study are unequal and that database attractiveness is not based on quality. As a result, global forest statistics providers may be motivated to put more effort into improving database attractiveness rather than quality, which is more challenging. Consequently, it is likely that the governance structure and strategies reported in these databases can substantially affect the reliability of numbers used in academic research and policy-decisions since they are generated from the related global forest statistics.