Turning a first-time visitor into a regular is one of the biggest wins for any business. Regulars spend more, recommend you to friends, and cost less to keep than new customers cost to find.
Harvard Business Review reports that increasing customer retention rates by 5% increases profits by 25% to 95%. Yet many first-time visitors never return, and only 30 to 40% of businesses report being committed to a lost-customer strategy.
Customers sometimes disappear even after a good experience. You need a way to find them and bring them back. Done well, that means better return on ad spend, higher customer lifetime value, and more revenue.
A cookie for the real world
Ecommerce has a head start here. Online stores re-engage past visitors with data from pixels and cookies: who visited and when, how long since the last visit, which products they viewed, and what they left in the cart.
There is no cookie in the real world. A physical business has to identify its lapsed customers first, then reach them with the right message.
In-store systems can fill that role by tracking visits and transactions. Point of sale (POS) systems have long served as customer databases, recording visits and purchase history. Knowing what customers spend helps you decide how much to invest in bringing them back.
Guest WiFi adds new contacts and records when customers visit. In exchange for WiFi access, customers opt in to receive marketing by entering an email address, which is linked to their device. When the customer returns with that device, the visit is recorded.
Together, POS and WiFi show who came in, when they lapsed, and how much they spent.
Target the customers you lost
Segmenting customers by visit history and lifetime value makes reactivation campaigns easier to manage. Pull a list of customers who have not visited in one to two times your average visit cycle, and you have a lapsed audience you can reach through digital advertising.
A West Coast coffee chain did exactly that. Because it could see how often each customer visited, it defined a lost customer as anyone who had not been in for 30 days or more, then built ad audiences from that list. It targeted only people within a few miles of each store and weighted budget by customer value: more spend on guests who had visited 10 times or more, less on guests who had visited two or three times. Its ad spend went to its most valuable lost customers near each location.
Use offers, and tier them
The offer matters too. The coffee chain chose not to include an incentive, and it did not need one. Its returning customers only needed a reminder.
Sometimes an offer is necessary, especially when a customer has been gone a long time: 80 to 90 days or more for many businesses. A good practice is to start with a small offer early, then move to a larger one to raise the odds of a return visit.
Paytronix, a loyalty platform, has shared recommendations from its own research: match the incentive to how long the customer has been away. Send the first offer at two to three times the average visit frequency. If guests usually visit every 30 days, send the first lapsed offer between 60 and 90 days of inactivity, valued at roughly 33 percent of average per-person spend. If guests typically spend $12, make the offer worth about $4.
Send a second, larger offer at four to five times the average visit frequency, worth 50 percent or more of average per-person spend.
Measure visits, not clicks
No win-back program works without measurement. For a physical business, the measure that matters is in-store visits and transactions, not clicks. Walk-Through Rate shows how many of the people who saw your ad actually walked into your location.
The bottom line
Winning back lost customers belongs in every physical business plan. Digital marketing used to make that hard for brick-and-mortar stores. With the lessons of ecommerce and the right in-store data, it is now practical. Focus on three things: customer data, messaging, and offers.
Originally published on the Zenreach blog on January 29, 2020. Zenreach is now Adentro. This article was lightly edited in October 2026.