This view turns the free-text comments customers leave when they rate their orders into a ranked list of frictions. Every rating comment is read and sorted into recurring issue themes, so you can see what customers are actually unhappy about, how widespread each issue is, and whether it is getting better or worse.
Each card is one issue
The large figure is that issue's share of comments in the current view. The right side shows growth versus the previous month (red means rising, green means falling).
Slice with the filters
By month, region (ROI 1 or ROI 2), city, vehicle category (2 Wheeler or Trucks) and sub category. Every number recomputes live.
Themes in the sidebar
Issues are grouped into pricing, behaviour, delivery, goods and app, so you can focus on one area at a time.
Read the source
Open the example comments on any card to see the real customer verbatims behind the number.
Based on 8,275 analysable comments across 40 cities (short and gibberish entries removed). Each comment was read and classified by a model rather than by keyword matching, and one comment can raise more than one issue, so the shares add up to more than 100%.