For the next installment of our Agency Interview Series, "Retention in the Age of AI," Gunjan, Flits' co-founder and CEO, sat down with Kurt Elster, founder of Ethercycle, a Shopify Plus Partner based in Chicago, Illinois, that's been helping ecommerce brands grow since 2009. Kurt's also the host of The Unofficial Shopify Podcast, which has racked up 500+ episodes and 2.5 million downloads of practical ecommerce advice, and he runs a YouTube channel with 235K subscribers.
If you've spent any time in Shopify circles, there's a good chance you've already heard his voice. What stood out in this conversation wasn't a slick framework or a new acronym. It was how unglamorous he thinks good retention actually is. Stay visible. Don't overbuild. Know your numbers before you spend money solving a problem you might not have.
Here's what he told us.
1. How do you typically approach retention for your clients, especially after the first purchase?
Kurt's starting point is always the same: figure out whether retention is actually a problem before you touch a loyalty program, an email flow, or anything else. His go-to metric is Shopify Analytics' Return Customer Rate, and he was specific about what good looks like, since it varies a lot by category.
- Subscription or consumable products (his example was coffee, a genuinely habit-forming one) can realistically push toward a 50% return rate. He framed that as close to the practical ceiling, even for the best-performing brands.
- Apparel is a different story. Hitting 25–30% RCR in that category would already be a standout result.
- Anything under 20% is a clear signal there's real room to improve.
He tied this back to the broader math every merchant is already running, whether they realize it or not: cost to acquire a customer, traffic, conversion rate, average order value. Repeat purchase behavior is what turns that acquisition math into actual lifetime value. Without it, you're just paying to reacquire the same growth over and over.
His actual tactical recommendation, once you know retention is worth pursuing, was almost anticlimactic: “stay top of mind”. That's it. Content generation is remarkably easy with AI now, and he was refreshingly candid that a lot of AI-assisted content lands somewhere in the middle of the quality curve, average and passable rather than exceptional.
His view is that sending out consistent, decent content beats sending nothing while waiting for perfect. Being forgotten costs more than being average.
A follow-up we couldn’t resist asking: which category should actually implement a loyalty program?
Kurt’s answer wasn’t about category at all. It was about purchase frequency. A customer who buys once a year isn’t going to be moved by a points balance; a loyalty program isn’t going to move the needle for that relationship. Consumable or subscription businesses are an easy yes. And any brand where customers are purchasing five or more times a year is a strong candidate, regardless of what they sell.
2. How are you seeing AI change the way brands handle retention and re-engagement right now? Any tools or approaches you're particularly excited about?
Kurt described himself as a heavy Claude Code user. He prefers working from the command line and directing AI from there rather than through a polished interface. He drew a line between power users who are comfortable operating in a terminal and everyone else, and argued we're heading into a world where anyone can build custom software. How well they build it just comes down to skill.
Applied to loyalty programs specifically, his point was that the platform matters less than most merchants think. Default configurations from loyalty apps typically don't fit a given brand out of the box. What actually determines success is how quickly a customer reaches their first reward, and whether the tier and reward progression logic makes sense for that specific business.
He said modeling that used to be back-of-napkin math and spreadsheets. Now, tools like Claude Code make it much faster to work through reward structures, something he now considers table stakes for a lot of brands. But the underlying point held: a loyalty program is a waste of time if a brand isn't willing to actually design a reward structure that works, AI-assisted or not.
A few other specifics he shared:
- Subscriptions are a strong retention lever, but really only make sense for consumable goods, harder to justify for anything else.
- Newsletter cadence: at minimum twice a month to stay visible.
On Shopify's broader AI direction, Kurt's read is that the company generally does one of two things with new technology: wait and see, or get out ahead and try to define the category.
With AI, he said Shopify chose the second path, pointing to early enthusiasm around GPT from Shopify co-founder and CEO Tobi Lütke, whose internal memo on AI adoption made waves across the industry, that's led to today's Universal Commerce Protocol, which he credits Shopify with leading.
On AEO (optimizing to get surfaced inside tools like ChatGPT), he was honest that the effort hasn't translated into meaningful additional sales yet. His honest read was that there's some payoff, but not much yet, and he admitted he doesn't know whether that changes long term.
He also pushed back a little on how many marketing agencies are selling AI right now. A lot of them are pitching agentic workflows and Slack bots meant to replace human roles, but in his experience those setups often add more complexity than they remove. His preferred approach is simpler: give the whole team shared access to one Claude account instead of layering in bots.
On Shopify Sidekick, the AI assistant built into the Shopify admin that can chat with merchants and take action on things like products and Flow automations, he said the adoption numbers are impressive, and his agency has gotten good feedback from clients using it.
Ethercycle is integrating Sidekick into their own free-gift-with-purchase app, and his personal workflow involves drafting a prompt in Claude, then pasting it into Sidekick to generate a Shopify Flow automation.
3. What kind of results have you helped clients achieve in terms of retention or repeat purchases? Any metrics or notable wins?
Kurt was straightforward here: he didn't have specific retention metrics on hand to share. What he offered instead was a look at the range of businesses he's worked with.
His agency work spans aftermarket automotive and a heavy amount of fashion and apparel, which he pointed out is a large vertical within Shopify generally, plus a handful of CPG and food and beverage brands, which he called an enjoyable surprise (a range not unlike what we've seen across our own merchant success stories). The one vertical he hasn't worked in yet but would like to: consumer electronics.
What This Tells Us
- Across the conversation, the theme that kept surfacing wasn't a new tool or tactic. It was discipline.
- Know your numbers before you build anything.
- Don't assume every brand needs a loyalty program.
- And if you do build one, put in the work to make the reward structure actually make sense, because the platform was never the hard part.
Are you a Shopify agency working on retention for your clients?
The Flits Agency Interview Series is open. If you're working on retention and have something worth saying, we're all ears.
