Uncertainties and biases in the measurements during a R&D process may lead to false conclusions and/or unnecessary repetition of costly experiments. A data quality evaluation and reconciliation process based on simple carbon mass balances calculated from ordinary measurements is proposed. The case-study was a database of 39 chemostats of polyhydroxyalkanoates (PHA) production by two
different bacterial strains with distinct bioreactor volumes and manufacturers. The assessment of this process showed that it can improve R&D efficiency by at least 15%, avoiding hours of extra experiments. The methodology is transferable to other databases unrelated to that of the case-study (PHA).