For the thermal analysis of structural or non‐structural timber members, using conventional simplified heat transfer models, thermal conductivity values of timber are normally calibrated to test results such that they implicitly take into account influences such as mass transport that are not included in the model. Various researchers and designers have used such effective thermal conductivity values, originally determined for standard fire exposure, to evaluate other fire scenarios such as natural fires. This paper discusses in qualitative terms some parameters that govern the burning of wood and their influence on effective conductivity values. Reviewing fire tests of timber slabs under natural fire conditions, the study explains why effective conductivity values, giving correct results for the ISO 834 standard fire scenario, should not be used in other fire scenarios. For this reason, the thermal properties of timber given in EN 1995‐1‐2 are limited to standard fire exposure. As shown by heat transfer calculations, the effective thermal conductivity of the char layer is strongly dependent on the charring rate and therefore varies during a natural fire scenario. It has also been shown that char oxidation during the decay phase in a natural fire has a significant influence on the temperature development in the timber member, since char surface temperatures exceed the gas temperature in the compartment or furnace. Using increased effective gas temperature as thermal action during the decay phase, and varying conductivity values for the char layer, fairly good agreement could be obtained regarding the temperature development in the timber member and the char depth. Copyright © 2005 John Wiley & Sons, Ltd.