Combatting illegal timber trade requires forensic tools that independently verify claimed geographic origin of timber. Chemical and genetic wood characteristics are potentially suitable tools, but their performance at small spatial scales is unknown. Here we test whether stable isotopes and microsatellites can differentiate Tali timber (Erythrophleum spp.) at the level of forest concessions. We collected 394 wood samples from 134 individuals in five concessions in Cameroon and Congo Republic. The nearest neighbour concessions were 14 km apart and the furthest pair 836 km apart. We constructed genetic profiles using eight nuclear microsatellite markers and measured concentrations of 18O, 15N and 13C. We differentiated provenances using PCA (microsatellites), ANOVA and kernel discriminant analysis (isotopes). Next, we performed assignment tests using blind samples (n = 12, microsatellites) and leave one out cross validation (LOOCV, isotopes). Isotopic composition varied strongly within concessions and only 13C differed significantly between two concessions. As a result, LOOCV performed only marginally better than random. Genetic differentiation among provenances was also relatively low, but private alleles were commonly found. Bayesian clustering analysis correctly assigned 92% of the blind samples, including those of nearby concessions. Thus, Tali timber can be successfully assigned to the concession of origin using genetic markers, but not using isotopic composition. Isotopic differentiation may be possible at larger spatial scales or with stronger climatic or topographic variation. Our study shows that genetic analyses can differentiate the geographic origin of tropical timber at the scale of forest concessions, demonstrating their potential as forensic tools to enforce timber trade legislation.