Latest / The Retention Marketing Podcast with Fexingo: Email, SMS, and Loyalty for Repeat Customers / How Branded SMS Chatbots Drive Repeat Purchases
Transcript
- Lucas: If today's episode sparks something you've actually used in your own marketing, the only way this show stays ad-free is listener support. And honestly, if today was worth a coffee to you, that's the link — buy me a coffee dot com slash fexingo. Luna: Yeah, we keep it simple. No tiers, no perks, just a way to say 'this was useful.' Lucas: Right. So — let's talk about SMS. Specifically, the branded SMS chatbot. Not the broadcast blast, not the human-staffed customer service line. The automated, two-way conversational flow that lives inside a text message thread. Luna: And this is different from the typical 'Hey, here's 20 percent off' text? Lucas: Completely different. Most brands treat SMS like a cheaper version of email — send a code, hope for a click. But the brands that are winning on retention think of SMS as a service channel first. They're building chatbots that feel like a helpful friend, not a discount cannon. Luna: Okay, give me a concrete example. Lucas: Let's use a pet-supply brand — call them 'Paw Provisions.' They sell premium dog food and treats, mostly subscription-based. But they noticed that a lot of first-time buyers never came back for a second order. So they built a three-message SMS chatbot triggered after the first purchase. Luna: Three messages total? Lucas: Three. Message one: shipping confirmation, but with a care tip. 'Your order of salmon kibble is on its way. Pro tip: transition your dog slowly over five days to avoid tummy trouble.' That's it. No coupon, no upsell. Just useful. Luna: And they automate that through a chatbot? Lucas: Yes. It's a branded SMS chatbot — the customer texts back a reply, and the bot handles it. Most people reply 'thanks' or a thumbs-up. The bot acknowledges it. Simple. Day seven, message two: 'How's your pup liking the food? Send us a photo!' The bot collects the photo and logs the reply sentiment. If a customer complains about pickiness, the bot offers a sample of another flavor. Luna: So they're gathering data and building a relationship in one message. Lucas: Exactly. Then day 21, message three: 'Looks like you might be running low on salmon kibble. Want to reorder your same bag? Reply YES and we'll send a link.' This is triggered by the actual order history — they know the bag size and the feeding schedule. The reply rate on that third message hit 42 percent. Luna: Forty-two percent reply rate on an SMS? That's huge. Lucas: It is. And here's the key metric: second-order rate within 90 days went up 34 percent for customers who received that flow versus those who got no post-purchase SMS at all. Luna: So the chatbot isn't just a gimmick — it's actually driving retention. Lucas: Right. And it works because it feels personal without being creepy. The brand isn't asking for personal details — they're using data they already have: purchase date, product, quantity. The bot asks for a photo, which is low friction and fun. Most people share a picture of their dog. Luna: And the care tip in the first message — that's a trust builder. It shows they know their product. Lucas: Exactly. The tip also reduces the chance the customer will have a bad experience. If they transition the food slowly, fewer digestive issues, fewer complaints, fewer returns. That's retention before the product even arrives. Luna: What about the brands that don't sell a replenishable product? A pet store works because dog food runs out. What if you sell something like apparel? Lucas: Good question. Apparel brands can still use SMS chatbots for post-purchase care. Message one: 'Your jacket shipped. Here's how to waterproof it for maximum life.' Message two: 'How does it fit? Reply with a thumbs up or down.' Message three, at day 30: 'We just dropped a new color. Members get first access — want a look?' That third message is a sales message, but it's earned because the first two were service. Luna: So the sales message only works if the utility messages come first. Lucas: Exactly. The brands that blast a promo code as the first text get high opt-out rates. The brands that lead with help get reply rates in the 30 to 50 percent range. Luna: I've seen a similar play from a skincare brand. They send a text after delivery: 'Your serum arrived. Start with every other night to avoid irritation.' Then day 10: 'How's your skin feeling? Reply with a word: GLOW or MEH.' If MEH, the bot offers a switch to a gentler formula. Lucas: That's smart — they're using the chatbot as a triage tool. A human doesn't have to answer every 'meh.' The bot handles the common issue, and only escalates if the customer says something like 'my face is burning.' Luna: And that saves the brand money on customer service while keeping the customer from churning. Lucas: Exactly. The chatbot cost them about 5 cents per message. A human agent costs a dollar or more per interaction. So if the bot can handle 80 percent of replies, the ROI is massive. Luna: What about opt-in rates? How do these brands get people to agree to SMS in the first place? Lucas: Most of them use a post-purchase modal on the checkout thank-you page: 'Get shipping updates by text.' That's a high-intent opt-in. The customer is already buying, so they want tracking info. Paw Provisions saw a 26 percent opt-in rate just from that prompt. Luna: And then once they opt in, they get the three-message flow. Lucas: Right. And here's the critical detail: the chatbot is branded. It uses the company name as the sender, not a short code that looks like spam. And the messages are written in the brand's voice — warm, specific, not robotic. Luna: So it's not 'You have a package en route.' It's 'Your pup's salmon kibble is on the truck, and here's how to make the switch easy.' Lucas: Exactly. The tone change alone can double reply rates. Luna: Any numbers on overall retention lift from an SMS chatbot program? Lucas: Industry benchmarks are still emerging, but the pet brand I mentioned saw customer lifetime value increase by 22 percent over six months for the SMS cohort versus the control. And the average order value on those third-message reorders was actually slightly higher than the first order — because the chatbot offered the same product, not a discount. Luna: So they're not training customers to wait for a coupon. Lucas: Exactly. That's the danger of blast SMS. You condition customers to only buy when you text them a deal. A well-designed chatbot conditions them to trust you for help. Luna: Is there a risk of over-messaging? Three messages over 21 days feels reasonable, but what if a brand tries to do more? Lucas: Big risk. The data shows that more than three messages in a post-purchase flow — unless the customer replies and asks for more — drives opt-out rates above 10 percent. Three is the sweet spot for most categories. Luna: And each message should have a clear purpose. Shipping and a tip. Check-in and a request. Then the reorder. Lucas: Right. No filler. Every message either helps, listens, or offers something relevant. Luna: What about the technology stack? Do you need a fancy platform to do this? Lucas: Not really. Most SMS marketing platforms — like Klaviyo, Postscript, or Attentive — now offer chatbot-like automation. You set up keyword triggers, conditional logic, and time delays. For a brand doing under a thousand orders a month, you can build this in an afternoon. Luna: And the key is mapping out the customer journey. What do they need at day one versus day seven versus day 21? Lucas: Exactly. The brands that succeed spend time writing the actual message copy. They test two or three versions of the care tip. They A/B test the timing. They look at reply rate, not just click rate. Luna: So the real differentiator is thinking of SMS as a conversation, not a broadcast. Lucas: That's the whole thesis. A broadcast says 'here's a deal.' A conversation says 'how's it going? need anything?' The chatbot makes that conversation scalable. Luna: I wonder if we'll see brands use AI to generate personalized care tips based on the customer's specific purchase. Lucas: Already happening. Some brands are starting to feed purchase data into a large language model to generate a unique tip for each order. 'Your order of organic chicken kibble for a golden retriever — here's a tip on portion sizes for active dogs.' That's where this is headed. Luna: But the core principle stays the same: serve first, sell second. Lucas: Exactly. And that's the takeaway for any brand, whether you sell dog food or software. The chatbot is just the tool. The strategy is generosity.