Thanks Chris. Answer is really useful.

Say your database handles <transactions> (highly frequent OLTP) of <shoppers> (frequent OLTP) on <outlets> (rare changes, single process). Say outlets are geolocated on an x,y grid, and you have frequent Queries searching for outlets nearby a geolocation. then you can benefit from from creating two bitmap indices on x and y when searching like

WHERE outlet.x BETWEEN x-d AND x+d 
AND outlet.y BETWEEN y-d AND y+d;

While this can also be done using geospatial indexing, the bitmap model can be extended to further attributes of outlets such as availability of product category a, parking b, child care c, etc.