A pharmacy or minimarket chain in Peru opens its second location and sets up what already worked at the first one: its own WhatsApp Business for orders and its own Yape account to collect payments. That makes sense location by location. The problem shows up when a customer who already bought at the first location starts buying at the second one too, closer to their office or new home. To the brand, it is the same buyer coming back. To each location, on its own, it is a brand-new customer who just walked in.
Each location builds its own record, with no idea the other one exists
One location’s WhatsApp Business does not talk to the other one’s. Neither do the two Yape accounts, nor the two spreadsheets where each manager logs their regulars. No tool is failing here: each one does its job well inside its own four walls. The problem is that the brand as a whole has no place where those spreadsheets meet, so nobody can answer whether the customer who bought yesterday in one neighborhood is the same one who buys every week in another.
This is different from a customer who pays by Yape at a single location, where the phone number already groups that repeating number. The phone number still works here, except the split is not inside one location, it is between locations that share no base at all.
What gets lost when a customer reads as two people
A customer who visits two locations of the same brand chose that brand twice, in two different places: a stronger loyalty signal than repeating at the same counter. Split into two separate records, that signal disappears. Each location reads it as an isolated purchase from someone brand-new, never as the fifth visit from its best customer.
How to start matching it without a new system
- Log the phone number the same way at every location. With or without the leading 9, with or without spaces: format drift is the first reason a manual match fails.
- Merge each location’s spreadsheet into one file, even weekly. No real-time system is needed to start: comparing phone numbers across locations and flagging the matches is enough.
- Prioritize whoever shows up at more than one location. That customer already proved something no single-location buyer did: the brand matters more to them than the convenience of one specific storefront.
- Only then evaluate a central system. A CRM shared across locations solves this permanently, but it is only worth it once the manual match confirms the problem is real and enough customers repeat across locations.
It is the same principle already used to retain a customer who pays by Yape inside a single location, applied now across locations of the same brand. It is also the work we do in data marketing in Peru: confirming whether the brand already has, without knowing it, a customer who chooses it more than any single location’s report can show.
If your business runs more than one location and you have never matched phone numbers across them, you probably already have customers from two locations that you are counting today as if they were from one.