A　large　volume　of　customer　reviews　is　generated　from　time　to　time　and　customer　requirements　are　presented　between　lines　of　online　opinions.　Many　studies　about　online　opinions　mainly　focus　on　the　extraction　of　customer　sentiment,　but　practical　concerns　regarding　the　integration　into　new　product　design　are　far　from　extensively　discussed.　To　enlighten　designers　about　how　consumers　differ　geographically　in　terms　of　their　preferences,　which　is　possessing　important　research　significance　and　practical　values,　is　not　well　investigated.　Specifically,　in　this　study,　online　reviews　are　invited　to　explore　market　regional　heterogeneity.　With　identified　product　feature　related　subjective　sentences　from　online　reviews,　a　straightforward　applied　approach　is　to　assume　the　ratio　of　the　number　of　satisfied　customers　to　the　total　number　of　customers　as　the　expected　percentage　of　satisfied　customers　across　different　regions.　However,　such　frequency　based　approach　becomes　unreliable　in　case　that　the　number　of　reviews　do　not　distribute　evenly.　Accordingly,　the　Bayesian　school　of　thought　is　utilized　in　which　statistics　of　data-rich　regions　are　invited　to　help　to　analyze　that　of　data-poor　regions.　Then,　a　hierarchical　Bayesian　model　is　proposed　and　it　assumes　that　the　expected　percentages　of　customer　satisfaction　in　different　regions　follow　a　certain　probability　distribution.　Finally,　taking　9541　mobile　phone　online　reviews　on　Amazon　as　an　example,　categories　of　experiments　were　conducted.　It　informs　the　significance　to　product　designers　about　the　value　of　online　concerns　on　analyzing　market　regional　heterogeneity　and　presents　the　effectiveness　of　the　proposed　approach　in　terms　of　discovering　customer　regional　differences.