•Dead wood varies greatly in quality as dependent on tree species and decay stage.•Differences in dead wood quality contribute greatly to forest invertebrate diversity.•Total invertebrate richness depends on specific tree species-decay stage combinations.•These findings can be used to optimize invertebrate diversity in managed forests. Dead wood availability and the variability in dead wood quality, i.e. tree species and decay stages, are often low in managed forests, which negatively affects biodiversity of invertebrate species. Leaving more (coarse) dead wood can increase invertebrate richness, but it remains unclear how many and which combinations of tree taxa and decay stages are required to optimize niche heterogeneity in managed forests. We investigated the diversity of the main arthropod groups associated with dead wood, i.e. millipedes, centipedes, isopods and beetles, through the first four years of decomposition of logs of twenty common temperate tree species placed in the “common garden” experiment LOGLIFE. We hypothesized that (1) invertebrate richness for combinations of a given number of tree species would be promoted by mixing both tree species and decay period and that (2) invertebrate richness increases up to a saturation point with more tree species at different decay stages added. We also hypothesized that (3) an increase in phylogenetic distance among the tree species in combinations would promote their overall invertebrate diversity. We found that the better combinations, in terms of invertebrate richness, after one and two years of decay, but not after four years, consisted of a mix of gymnosperms and angiosperms, indicating that variation in tree species is especially important during the initial decomposition period. The best combinations in terms of invertebrate richness consisted of at least one tree species from each decay period, indicating that also variation in the decay stage of the tree is important to promote invertebrate diversity. We observed that at least four wood types were required to approach the 95% saturation point for species richness. The third hypothesis, that dissimilarity in phylogenetic position could be a predictive tool for increasing invertebrate richness in combinations of tree species, was not supported by our results. Thus, in order to maintain diversity of dead wood invertebrates in forests we recommend not only to provide richness in tree species, but also to plant particular combinations of trees (preferably angiosperm-gymnosperm combinations) that differ in the invertebrate communities they typically host and to temporally spread the logging of trees. This way the logging residues cover different resources and habitats at each moment in time, which is likely to result in a large diversity of dead wood invertebrates.