The dataset consists of two data tables in comma-separated values (.csv) format, “Faecal_microbiota_counts.csv” and “OTU_names.csv”. Faecal_microbiota_counts.csv contains the community matrix of counts of OTUs from 10 ponies sampled on high starch and high fibre diets. Column 1 contains the row names, detailing pony id number, the sampling time in the cross-over design and the diet type. The subsequent four columns describe the information contained in the row names. Column 1: pony identification code. Column 2: diet code. Column 3: sampling period in cross-over experimental design. Column 4: dummy variable describing the direction of diet change in the cross-over design. Columns 5-2057: OTU identification number. OTU_names.csv contains the taxonomic names for each OTU. Column 1: taxonomic level - kingdom. Columns 2: taxonomic level - phylum. Column 3: taxonomic level - class. column 4: taxonomic level - order. Column 5: taxonomic level - family. Column 6: taxonomic level - genus. Column 7: taxonomic level - species. Column 8: OTU identification number.