Loss on ignition (LOI) is one of the most widely used methods for measuring organic matter content in soils but does not have a universal standard protocol. A large number of factors may influence its accuracy, such as furnace type, sample mass, duration and temperature of ignition and clay content of samples. We conducted a series of experiments to quantify these effects, which enabled us to derive (i) guidelines for ignition conditions (sample mass, duration and temperature), (ii) temperature-specific soil organic matter (SOM) to soil organic carbon (SOC) conversion factors and (iii) clay content-dependent correction factors for structural water loss (SWL). Bulk samples of a sandy soil (4% clay) and a silt loam soil (25% clay) were used to evaluate the effects of ignition conditions. Samples with a range of clay contents (0-50%) were used to quantify conversion and correction factors. Two furnaces, one without and one with pre-heated air, did not show significant differences in terms of within-batch LOI variability. In both furnaces less combustion occurred close to the door, which necessitated tray turning at half-time as this reduced the standard deviation per batch significantly. Variation in mass loss declined exponentially with sample mass (range, 0.15-20g). The LOI increased with duration at lower temperatures (≤550°C) for the sandy soil. At greater temperatures (600 and 650°C), no effect of duration was found. For the silt loam soil, LOI values increased with duration for each temperature, which was attributed to SWL. The SOM to SOC conversion factor decreased strongly with temperature at an ignition duration of 3hours from 0.70 (350°C) to 0.57 (500°C) and stabilized around 0.55 between 550 and 650°C, indicating that at temperatures ≥550°C all SOM had been removed. The clay correction factor for SWL increased from 0.01 to 0.09 as the temperature of ignition increased from 350 to 650°C. To minimize within-batch LOI variation we recommend a standard ignition duration of 3hours, tray turning at half-time, a sample mass ≥20g and temperatures equal to or greater than 550°C. To avoid over-estimates of SOM through structural water loss, the presented SWL correction procedure should always be applied.