Yeah, the old index format had many factors contributing to it taking up considerable space. As a single data point, we have a beer-search sample app, this includes a data directory with 29MB of JSON files. In the old index format, with mapping that did a realistic configuration of indexing many fields, and storing some of them, the bleve index size was over 200MB.

With scorch the index size is 22MB. Query performance is comparable (and we haven't even gotten to really tuning this yet).