Latest / The Retention Marketing Podcast with Fexingo: Email, SMS, and Loyalty for Repeat Customers / How a Coffee Chain Uses SMS to Retain Through Predictive Reordering
Transcript
- Lucas: So there's this regional coffee chain in the Pacific Northwest, about forty locations, and they've been quietly using SMS to predict when customers are about to run out of their favorite whole-bean blend and send a text before the customer even thinks to reorder. Luna: Wait, so they're texting people 'you're probably out of coffee'? That feels a little bit like a friend who knows your fridge better than you do. Lucas: Exactly. And apparently it's working. Their repeat purchase rate within thirty days jumped about eighteen percent for customers who opted into these predictive texts. But what's interesting is the mechanics. They're not just looking at past order dates. They're layering in purchase intervals, average consumption per household, and even local weather data. Luna: Weather data? How does rain affect coffee consumption? Lucas: That's the clever part. On cold, rainy days, people tend to brew more at home. So their model adjusts the predicted 'run-out date' based on a ten-day forecast. If a customer typically buys a twelve-ounce bag every two weeks, but there's a week of heavy rain coming, the system might push the reorder text a couple of days earlier. Luna: So they're basically forecasting consumption like a utility company forecasts energy demand. That's way more sophisticated than a simple 'hey, you haven't ordered in a while' message. Lucas: Right. And the tone of the texts matters just as much as the timing. They're not blasting 'LAST CHANCE' or 'HURRY'. The message is something like 'Hey, looks like you might be running low on the House Blend. Want to pickup a bag this week? We'll have it ready.' Casual, helpful, like a barista who remembers your order. Luna: That's the key. It's not a sales pitch, it's a reminder. And they must have really good data on consumption patterns to make that work without being creepy. Lucas: They do. And they're careful about frequency. One text per cycle, max. If the customer doesn't respond, they don't follow up with another one. It's a single, well-timed nudge. And they're also using purchase data to segment: someone who buys decaf every Monday gets a different message than someone who buys a variety pack monthly. Luna: So what's the biggest challenge for a smaller brand trying to do this? I imagine you don't need a data science team, but you do need some kind of logic. Lucas: Exactly. A lot of retention platforms now offer predictive triggers out of the box. You set rules like 'if a customer's average interval between purchases is X days, and it's been X minus two days since their last purchase, send a text.' The weather layer is a nice to have, but even without it, you can start with simple interval-based predictions and refine from there. Luna: And what about opt-in? Because SMS is so personal, getting consent is non-negotiable. Lucas: Absolutely. This chain only texts customers who've explicitly opted in, usually at the point of sale or during checkout online. And they make it easy to opt out, which builds trust. They also track delivery and click-through rates obsessively, so they know which messages resonate and which feel like noise. Luna: That makes sense. And it's a good example of how retention marketing is moving from reactive to proactive. You're not waiting for a customer to lapse; you're anticipating their needs before they even realize they have them. Lucas: Right. And that proactive approach is what separates a loyalty program that feels like a points ledger from one that feels like a relationship. This chain isn't just offering a discount on their next bag; they're showing they understand their customers' habits. Luna: You know, this whole conversation reminds me of something I've been thinking about with our own show. We spend so much time each week talking about these marketing strategies, and I've actually started applying some of the ideas to how I run my side project's email list. Just simple things like segmenting and timing better. Lucas: That's great to hear. And I think that's part of why we keep doing this show, honestly. When these conversations spark something that you actually use, that's the best feedback we can get. It's also why we keep the show ad-free — it just feels more honest when we're recommending things purely on merit. Luna: Yeah, and for anyone who's found value in these episodes, there's a simple way to keep us going: buy me a coffee dot com slash fexingo. It's a small gesture, but it genuinely helps us cover the costs of producing the show. Lucas: Right. And no pressure at all. Even just knowing that people are listening and applying these ideas is what keeps us motivated. But if you do want to support, that's where to do it. Luna: So, back to predictive SMS. What about the privacy angle? Are customers okay with a brand knowing exactly when they're likely to run out of coffee? Lucas: That's the big question, and it's a fine line. The chain found that as long as the message is framed as helpful and the customer retains control, they're actually happy to receive it. They surveyed their opt-in customers and over seventy percent said they appreciated the reminders. It's when brands get too aggressive or start predicting things that feel intrusive that it backfires. Luna: So the key is to keep the prediction about the product, not about the person's life. 'You might be out of coffee' is fine, but 'You're probably stressed and need caffeine' would be a step too far. Lucas: Exactly. And they also give customers the ability to adjust the frequency or pause the texts entirely. That level of control builds trust. Plus, they make it dead simple to opt back in if someone changes their mind. Luna: For a small business without a data science team, what's the simplest version of this? Let's say someone runs a boutique skincare brand. Lucas: Start with the product's lifecycle. If a serum lasts about four to six weeks with daily use, and the customer has bought it twice at five-week intervals, you can set a trigger at week four to send a text: 'Your favorite serum might be running low. Reorder today and get free shipping.' That's a simple rule-based prediction that doesn't require machine learning. Luna: And you can test and refine based on response rates. If people are reacting well, you can layer in more variables like seasonality or purchase history. Lucas: Right. And the beauty of SMS is that it's immediate. Email can sit in an inbox for days, but a text is typically read within minutes. So when you time it right, the impact is much stronger. Just remember to always tie it back to value for the customer, not just your bottom line. Luna: I love that. It's a good reminder that retention is ultimately about being helpful, not being clever with data. So what's the takeaway for our listeners today? Lucas: I'd say this: stop waiting for customers to lapse. Look at your purchase data and find the natural rhythm of your product. Then use SMS to meet them at that moment with a helpful nudge. It's a low-cost, high-impact retention strategy that's still underused. Luna: And if you've got a story about a brand that does this well, we'd love to hear it. Maybe it'll be a future episode. Lucas: Absolutely. Until next time, keep your customers close and your texts timely.