Trees exhibit different growth rates and timings of wood formation. However, the factors explaining these differences remain undetermined, making samplings and estimations of the growth dynamics a complicated task based on technical rather than statistical reasons. We collected weekly wood microcores in 159 balsam firs (Abies balsamea (L.) Mill.) from April to October 2018. We tested spatial autocorrelation, tree size, and cell production rates as explanatory variables of xylem phenology, and we assessed the margin of error for different sample sizes. Xylem growth lasted between 40 and 110 days, producing between 12 and 93 cells. Neither the spatial proximity nor the size of individuals explained the variability in xylogenesis. A higher cell production corresponded to an earlier onset and later ending of xylem differentiation. A sample size of 23 trees estimated xylem phenology at 95% confidence level with a margin of error of a week. The relationship observed between the timings of xylem differentiation and annual cell production support the hypothesis of a potential connection between xylogenesis and carbon uptake in trees. When studying growth dynamics at high temporal resolutions, sample size assessment must suitably represent the variability in xylem phenology and consider the annual growth rates of the trees. The possibility of increasing sample size over the growing season can help evaluate the enhancing variability among trees. Sampling may profit trees located in the same plot, given that the variability has a similar magnitude within and among plots. Incorporating the variability in phenology and growth rates among individuals could allow more reliable upscaling of carbon allocation at stand or landscape level.