Latest / The Retention Marketing Podcast with Fexingo: Email, SMS, and Loyalty for Repeat Customers / How a Meal Kit Brand Uses Email to Retain Subscribers Through Recipe Personalization
Transcript
- Lucas: So there's this meal kit company that figured out something pretty clever about retention. It's not just about sending a weekly recipe email. It's about making that recipe feel like it was chosen specifically for you, based on what you've actually cooked before, not just what you said you liked. Luna: Yeah, I've seen those emails. They always show a 'chef's pick' or a 'customer favorite.' But how do they actually personalize it beyond just segmenting by vegetarian or not? Lucas: Right, that's the surface level. What this brand does is track every single recipe you've ordered, plus your ratings, plus what you've skipped, and even how long you take to decide. Then their algorithm builds a taste profile. So if you've ordered three different Thai curries and rated them all five stars, the next email's top pick is probably a new Thai-inspired dish. And they send it with a subject line like 'A new curry for your collection.' Luna: That is so specific. I can see how that would feel less like spam and more like a friend recommending something. But does it actually move the needle on retention? Or is it just a nice touch? Lucas: They shared some numbers at a conference last year. Subscribers who engaged with those personalized recommendations had a churn rate that was roughly thirty percent lower than those who didn't. And the open rates on those emails were above fifty percent, which is huge for email. Luna: Thirty percent lower churn, that's not nothing. But how do they keep it from feeling too robotic? I mean, algorithm-driven emails can sometimes feel a bit cold. Lucas: That's exactly what they focused on. They added a 'chef's note' — a short, handwritten-style paragraph about why this recipe might be a good fit, referencing a past order. Like, 'Since you loved our basil chicken, we thought you'd enjoy this basil tofu with crispy rice.' It's simple, but it makes the email feel human. Luna: So it's a blend of data and storytelling. That's a good balance. And I'm guessing they also use the feedback loop, like asking you to rate the recipe after you cook it? That would feed back into the algorithm. Lucas: Exactly. They ask for a rating after every delivery, and they also ask a quick follow-up: 'Would you cook this again?' That single yes or no question is gold for their model. It's not just about liking a dish; it's about whether you'd repeat it, which is a stronger signal for retention. Luna: That makes sense. And then they can use that data to avoid recommending something similar next time if you said no. It's like they're learning from every interaction. Lucas: And that's the thing that struck me. Most brands treat email as a one-way broadcast. This brand treats it as a conversation. Every click, every open, every rating informs the next email. They even track whether you clicked the recipe link but didn't order — that's a signal they can follow up on with a gentle nudge. Luna: A follow-up nudge? Like what? Lucas: If you opened the email, clicked the recipe, but didn't add it to your box by Thursday, they might send a short reminder on Friday: 'Still thinking about that pad thai? It's available for next week's box.' It's timely, relevant, and not pushy. Luna: That is smart. It's catching that moment of indecision and giving a little push. And it's all automated, right? I mean, with that volume, it has to be. Lucas: Fully automated. But the copy is written to sound like a human wrote it. They use natural language, contractions, even emojis occasionally. And they always give you the option to say 'not interested' — which also feeds the model. Luna: So it's a constant feedback loop. That's pretty sophisticated for a meal kit brand. Are there any pitfalls? Like, what if someone gets a recommendation they don't like and it feels off? Lucas: They've built in a fallback. If the algorithm is unsure, they default to a 'popular in your area' pick. And they always include three other recipe options, so the personalized one is just the top choice, not the only one. That reduces the risk of feeling like the brand doesn't know you. Luna: That's a good safety net. And I'm guessing they also let you manually adjust preferences, like 'I'm trying to eat more plant-based' or 'I'm allergic to nuts.' Lucas: Absolutely. They have a preference center in the email footer where you can update dietary restrictions, cooking time preferences, even spice tolerance. And that data gets merged with the behavioral data. So it's a combination of explicit and implicit signals. Luna: So the email becomes a dynamic, living profile of the customer. That's a lot more engaging than a static newsletter. I bet their retention numbers reflect that. Lucas: They do. They saw a fifteen percent increase in subscriber lifetime value within six months of rolling out the full personalization engine. And their reactivation campaigns, for people who had paused or canceled, also improved — they could send a 'we miss you' email with a personalized offer based on past favorites. Luna: Reactivating lapsed customers is a whole other beast. But if you can say, 'Remember how much you loved our shrimp tacos? They're back for one week,' that's a powerful hook. Lucas: Exactly. And they tie that to a limited-time availability, which creates urgency without being pushy. That's a classic retention tactic — using scarcity and nostalgia together. Luna: Before we go further, I have to say, these conversations we have on the show, they've genuinely sparked ideas for my own side projects. And I know we're not the only ones who feel that way. A lot of listeners have told us they've applied things they heard here. It makes me think about how we keep this going. Lucas: Yeah, it's a two-way street. And honestly, the support from listeners is what keeps this ad-free and independent. A couple of dollars a month is genuinely what keeps these going — buy me a coffee dot com slash fexingo, if you've gotten something out of them. Luna: Yeah, it's like the personalization we're talking about — a small gesture that goes a long way. So, back to the meal kit brand, are there any lessons for smaller brands that don't have a huge data team? Lucas: Great question. You don't need a massive algorithm to start. You can begin with simple rules. For example, if a customer has ordered the same product category twice in a row, show them more of that category in your next email. That's a manual version of personalization. Luna: And you can use surveys or preference centers from day one. That doesn't require a data scientist. Just a simple 'tell us what you like' form. Lucas: Exactly. And then layer in the behavioral data as you go. The key is to start with a clear hypothesis — like 'customers who order the same cuisine twice are more likely to order it again' — and test it. That's the same methodology, just scaled down. Luna: So the core principle is: use every interaction to learn and adapt. It's not about being perfect out of the gate; it's about building a system that gets smarter over time. Lucas: That's the takeaway. And it's why I find this case so compelling — it's not a flashy campaign, it's a systematic approach to retention. And it's something any brand can start doing today, even with a spreadsheet and a simple email tool. Luna: Spreadsheet and a simple email tool, that's actually reassuring. I was worried it would take a whole tech stack. Lucas: Nope. The tech is just an enabler. The real magic is in the strategy and the copy. If you can write an email that feels personal, that's more important than any algorithm. Luna: And that's a skill anyone can work on. So, for our listeners, maybe the takeaway is to look at your current email list and think: are we sending the same message to everyone? Or are we actually listening? Lucas: That's a perfect way to put it. Are we listening? And that's the question that separates retention marketing from just sending emails. I know we'll be thinking about that for a while. Luna: Definitely. And maybe next time we'll talk about how to handle the data privacy side of all this personalization. Lucas: That's a great topic for a future episode. For now, thanks for joining us, and we'll catch you next time.