While applying a segment to the report an error comes up "One of the segment or the search in the visualization contains a text search that returned too many results"
When a segment is applied, our servers convert the segment criteria from a list of strings to a list of hashes matching those strings. If one (or more) criteria in the segment match a large number of hashes then the request becomes too large and thus generates the error.
To improve performance and make everything uniformly sized, every variable value has a corresponding 32/64-bit hash. All reports run on these hashed values until the data is presented, where each hash is replaced by the original text.
The usual way to combat this is to simplify the segment, convert some of the criteria to use classifications, and/or use a smaller date range for the report (since the hash tables are arranged by month, each new month adds the potential of more hashes matching).