This week, I had Nate Okonkwo from Remarkable AI on the Retention Edge podcast, talking AI, personalization, and winning back lapsed customers.
One number that came up in the conversation really stood out:
$0.11 - that's the industry average revenue per email for post-purchase win-back campaigns. Eleven cents per send. For lapsed customers six months out or more.
Now $0.11 isn't zero. Send enough emails and the math still adds up. But it's the kind of number that tells you nobody's really trying that hard.
Nate's brands are pulling $5 to $17 per email on the same audience.
That's a 50x lift. With no discounts attached.
Once you understand why most brands are stuck at 11 cents, the gap actually makes sense.
The gold mine nobody is mining
Pick any DTC brand that's been around for five-plus years. They have a list. Tens of thousands, sometimes hundreds of thousands of customers who bought from them once and never came back.
Most of those customers didn't have a grand, well thought-out reason for leaving. They’re not making a statement against the brand.
They got busy. Forgot the brand existed. Got distracted by the next thing.
You already paid to acquire every one of them. You already have their email. You already know what they bought, when they bought it, and what they were browsing the day before they checked out.
That sits there. For most brands, indefinitely.
Nate told me about a customer of theirs who'd been in business for years and had hundreds of thousands of one-time buyers. She'd never set up a system to bring any of them back.
It’s not that she didn’t care. Just no one on their team had time, and the tools they used weren’t doing the job.
It's the same story at almost every brand I talk to.
Why most win-back systems are broken
There are three versions of the problem, and most brands are guilty of at least one.
The first: most brands aren't running win-back campaigns at all. No flow. No automation, no tracking. The customer hasn't ordered in a year and the brand has no system that triggers on it. That's the majority.
The second: the flow exists, but it's stale. It was built five, eight, sometimes ten years ago in Klaviyo or whatever email tool was hot at the time. The marketer who set it up has long since left the company. The flow is still pushing products that don't exist on the website anymore. Nobody touches it because nobody owns it. The numbers haven't moved since 2019.
The third: the message is generic. "We miss you. Here's 10% off." Same email, discount, subject line as every other brand on the planet. The email lands in the promotions tab, and the customer doesn’t even see it.
There's a structural reason underneath all three. Legacy email automation tools were built for a human to log in and configure. They assumed every brand would have a CRM lead with the time to maintain a thirty-step lifecycle program.
Look at any DTC brand right now. The marketing team is half the size it was two years ago. They're managing a customer base that's twice as big. They're firefighting Mother's Day campaigns and BFCM planning. Nobody on that team has the time to go back and rebuild the win-back flow that was set up before they joined the company.
So it sits there producing 11 cents per customer.
What high-performing win-back campaigns look like
The blueprint behind a successful win-back system is creating campaigns and emails that look almost nothing like a typical "we miss you" blast.
Take Melinda Maria, a jewelry brand Nate’s company works with. The brand has a lifetime guarantee on every piece, which most customers don't realize when they buy. So the win-back message is something like, "Hey, how's that necklace working out for you?"
That's it. Not even a discount.
What happens is interesting. Most people don't reply. They just go back to the site and buy something else.
Some people do reply, sometimes saying things like, "Honestly, one of the stones fell out and I haven't worn it in months." The brand replaces the piece for free. The customer is shocked a brand actually cares. Then they buy again.
That brand is generating $17 per email - with no discount.
Here’s a deeper breakdown of the mechanics behind the most successful win-back flows:
Real context, not segments. Not starting with "lapsed buyers in the $50-$100 AOV cohort." Starting with order history, browse behavior, support tickets. The richest signal in your CRM is usually the support log, and almost nobody is reading it back into their email program. If a customer asked about sizing two months ago, your win-back email should reference that. If they returned an item, your win-back email should not push them the same item again. Most brands are blasting customers about products they literally just bought.
The next product in the routine. A beauty retailer Remarkable worked with wanted to push haircare, skincare, and bodycare, but not fragrance or makeup. The system routes each customer to the next-best product in their actual routine. Bought a cleanser? Next email recommends a sunscreen. Not a fragrance launch, not the trending makeup collab, not whatever's on sale that week.
No discount. This is the part that breaks people's brains. The math is so much better when the email is genuinely helpful that you don't need a 10% off lever to pull. Most brands have been training their list for years to wait for the discount. Melinda Maria's flow does the opposite, and they're at $17 a send.
The luxury world has been pulling these numbers for decades, doing it by hand. Personal stylists, real follow-ups, "how did you enjoy the piece" outreach from someone in the store. It works. It's never scaled below the luxury tier because no one had the time to do it for 100,000 customers.
That's the actual unlock here. AI doesn't make the email smarter. It makes the human-feeling email scalable.
The math
When you run the numbers, you see how much upside there is to fixing your win-back flows.
A beauty retailer ran a pilot with Nate’s company. The target was $20,000 in revenue from reaching out to 4,000 lapsed customers. They hit $26,000 from 1,500 customers. Less than half the audience, but more than the goal.
Versus a control group, the top campaigns drove a 3x lift.
Across their book, the average is $5 to $17 per email. Industry baseline is $0.11.
Even at the low end of that range, you're at 45x. At $17, you're at 154x.
Round it down to 50x and it's still the highest-leverage change you can make to your retention program this quarter. Probably this year.
What I'd take from this
A few things from the conversation that apply broadly, whether or not you ever talk to Remarkable:
Audit your lifecycle flows by date. When were they last touched? If the answer is "before 2023," the flow isn't optimized. It's archived. Treat it that way.
Pull the support log into the same brain as the email tool. The reason "how's it working out" lands for Melinda Maria is because there's actually a brand on the other side that responds when a stone falls out. Service-led messaging only works if there's actually service behind it.
Drop discounts from the win-back path entirely, then measure. Easiest A/B test in retention. Almost nobody runs it. Genuine, contextual outreach often outperforms a 10% off code, and it doesn't condition your customers to wait for the next sale.
Stop blasting customers about products they literally just bought. Sounds obvious. Open your own brand's automation flows right now and check.
AI can do some amazing things for you in terms of personalization, but the win-back gap isn't really about AI. It's about the fact that most brands have been treating their list like a list when they should've been treating it like people. AI is what makes that scalable at six-figure customer counts.
The brands that get this right will outpace their categories on profit per customer for years.
Check out the full episode with Nate below:
If you want to hear more, or work with him, find him on LinkedIn or head to beremarkable.ai.
A Better Way to Launch a Custom App
In 2026, just about every brand needs a mobile app. And they need a mobile app that actually matches their brand - and doesn’t take a team of six to run.
MobiLoud is the smartest way for mid-market and enterprise brands to launch a mobile app. The app is powered by your existing web stack, which means no rebuilding, no separate codebase to manage.
You get everything you need in your app, including custom app pages and native push notifications, with all the design and content managed through your existing systems.
Brands like Jack & Jones, John Varvatos, Cold Culture and Estee Lauder use this approach to have high-quality, professional apps that drive millions in revenue, with minimal management required.
Want to see how MobiLoud can help you launch a low-overhead, custom mobile app? Get a free preview of your app or DM me on LinkedIn.
Quick Hits
Amazon disclosed at its April earnings that monthly active users on Rufus are up 115% YoY, with engagement up 400%. The number that matters: customers using Rufus are 60% more likely to complete a purchase, and the assistant generated nearly $12B in incremental annualized sales in 2025. Whatever you think of agentic commerce, the on-Amazon version is no longer hypothetical.
Over 500 shopping apps were added to ChatGPT in May alone, including from Target, Walmart, Sephora, Starbucks, Instacart, and Uber. The catch, per multiple analysts in the piece: there's no visible success story yet. Adoption and conversion are reportedly low, and ChatGPT's own Instant Checkout was quietly retired after failing to drive sales. Most of the rush feels like metaverse-cycle reflex - launch because the technology is trendy, not because the use case is solved.
Etsy rolled out a native ChatGPT app this week. Tag @Etsy in a prompt ("help me find a Mother's Day gift under $100 for my mom who loves gardening") and it surfaces listings from Etsy's 100M-item catalog. This is Etsy's second swing at ChatGPT after participating in the now-retired Instant Checkout. Other retailers in the same wave: Angi, SeatGeek, Tubi, Wix.
OpenAI dropped the $50K minimum on its self-serve ads manager last week. SMBs, startups, and global brands can now buy ChatGPT ads directly. CPA bidding and third-party measurement are "in motion" but unscheduled, and the only ad format right now is a favicon with text. Worth watching, not worth shifting budget yet, but another reminder of how much is being built on top of a platform that didn't exist as an ad channel six months ago.
Amazon opened up its freight, distribution, fulfillment, and parcel network to any business, not just Amazon sellers. P&G, 3M, Lands' End, and American Eagle are early customers. Amazon is positioning this the way it positioned AWS: decades of internal infrastructure exposed as a service. If it works, it puts every 3PL in the country in the same position the data center business was in 15 years ago. And it gives DTC brands a direct route to Amazon-grade logistics without ever touching the marketplace.
Recharge announced the acquisition of Skio in April. Together the two platforms power 20,000+ brands and process around $20B in annual GMV, the bulk of Shopify's subscription volume. Both will keep operating independently for now, with the combined roadmap rolling out over the coming months. The two biggest subscription stacks on Shopify are now under one roof, which says something about where the value is in this market: not in acquiring the next first-time buyer, but in building the infrastructure that makes the second, fifth, and twentieth purchase happen on autopilot.
That’s all for this week.
If you have any thoughts on what we discussed here, or anything else you’d like to see covered in relation to retention, CX and growth, just hit reply and let me know.
Until next week,
Pietro and The Retention Edge Team
PS: go to our website to get a preview of your app for free, or shoot me a DM on LinkedIn to talk about it.