There are increasing tools available that can rigorously investigate complex demographic histories of natural populations from genomic data. This includes the methods for evaluating genotype uncertainty from common sequencing technologies and models for testing hypotheses of gene flow or population size change. However, most method development occurs with haploids or diploids in mind. Understanding the population genetic processes of polyploid or mixed ploidy systems can be beneficial for speciation genomic investigations, resolving taxonomy of difficult complexes, and potentially the conservation policy decisions based on polyploid plants. Here, strategies for accommodating polyploids in population genomic analyses are explored with two datasets, one target-enrichment and one ddRADseq, from Loudetia simplex (Tristachyideae: Panicoideae: Poaceae) sampled across the Central Highlands of Madagascar. Loudetia simplex is a tussock-forming C4 grass that is ecologically dominant in much of Madagascar’s Central Highland grasslands and can be found in other mainland African grasslands. However, demographic investigations have largely been hampered by uncertainty in life history and the presence of both diploid and triploid populations in Madagascar. Comparisons are made between genotyping methods that directly account for the correct allelic dosage and those that ignore the allelic dosage of polyploids. Practical guidance on genotyping in mixed ploidy systems is provided. For cases where accurate genotype calls are a concern, analyses of population structure should largely be robust to the loss of information from ignoring allelic dosage. Findings support ancient origins of Madagascar’s grasslands, and that these grasslands experienced population size expansions through the Late Pleistocene, coincident with paleoclimate change. Although there are regions of Madagascar where deforestation and biodiversity loss are a serious concern, our results bolster a nuanced view that much of central Madagascar prior to human colonization was a mosaic of open and wooded grassland. Combined with a rich cross-disciplinary history in the investigation of Madagascar’s open biomes, these population genomic investigations can provide guidance on which grasslands should be protected and which likely represent recent anthropogenic expansions.