Although Europe, Canada and the US have switched from cellular to acellular pertussis vaccines, most developing countries will continue to use the more cost effective cellular vaccine. Consistency of production however is the typical problem inherent to cellular vaccines. Optimising the production process of cellular pertussis bulk suspensions using product potency as a measure is not possible, since the mandatory animal test to measure potency has little discriminatory power. To circumvent this problem, this study focussed on measuring process parameters related to consistency and potency instead, even though the extent of those relationships could not be quantified. Critical evaluation and modification of individual process steps lead to 2 optimised production processes, NVP-96 and NVP-THIJS. These were compared to the original NVP production process in terms of antigen and biomass content, potency, toxicity and immunogenicity in mice. The batch to batch variation for both optimised products was clearly less than the original product for all parameters tested. The biomass content of the NVP-THIJS product was 15% lower than that of the NVP-96 product, while the immunogenicity in mice was twofold to threefold higher. The stability of the NVP-THIJS product remained higher than the NVP-96 product over a period of 2 years, while the decline of the potency of both suspensions was comparable.