eCommerce Podcast

Up to 40% of Amazon US sales now touch Rufus, Amazon's AI shopping assistant — and most sellers haven't optimised a single product for it. That's the gap David Balan, founder of Selluna, walked through with Matt this week. The customer who used to type "blue football" into a search bar is now asking a chatbot what kind of football would suit a six-year-old, and the products that show up first are the ones answering the right 25 questions.
David has spent the last few years helping thousands of brands across the US, UK, Europe and Asia-Pacific sell more on Amazon. In this conversation he laid out a three-pillar framework — traffic, clicks, conversion — and made the case that anyone selling on Amazon without a Rufus strategy right now is leaving 12-24 months of compounding advantage on the table.
Key Point Timestamps
[04:29] Why the Amazon algorithm is simpler than most sellers think
[05:28] The three pillars of Amazon — traffic, clicks, conversion
[07:00] Rufus, the top 25 questions, and how to answer them in priority order
[12:03] Traditional Amazon SEO — 80+ keywords, frequency, and the "stuffing" myth
[19:30] Why Rufus optimisation now is the new 2016 SEO play
[24:57] Why not A/B testing the main image is "criminal" in 2026
[33:08] The conversion stack — seven images, A+ Content, brand story, premium branding
[39:32] Text speaks to the algorithm, images speak to the customer

The Algorithm Is Simpler Than Sellers Want To Admit [04:29]
Most sellers treat Amazon's algorithm like an opaque, mystical force. David disagrees. "The algorithm is more simple than we like to make it," he told Matt. "The algorithm is trying to make Amazon as much money as they can get out of it. So once we understand their goal, if we align our goals with theirs, then everyone's going to be happy and you're going to be selling more at the end of the day."
That's the philosophical shift. Stop trying to outsmart Amazon. Start asking what Amazon wants and give it more of that. Every metric David teaches — traffic, clicks, conversion — is a proxy for the same goal, putting the right product in front of the right customer at the right time.
The Top 25 Questions That Make Rufus Recommend Your Product [07:00]
Rufus is Amazon's ChatGPT-style assistant, embedded in the marketplace. Instead of typing "football" into a search bar, customers now ask questions — what colour, what size, what age group, premium or budget. Rufus interrogates their needs and recommends specific products. Some data suggests up to 40% of Amazon US sales now involve Rufus at some point in the journey. David thinks the real number is closer to 20%, but it's growing fast and rolling out across Europe right now.
The play is simple. Rufus, like any AI, leans toward whatever it has the most data on. To make a product the answer, sellers need to identify the top 25 questions customers ask about their category and then answer those questions across every part of the listing in priority order.
The most important questions go in the title
Next most important in the bullet points
Then the product images
Then the back-end attributes
Then the reviews and Q&A

Selluna's free tool at app.selluna.ai/audit will surface the top 25 questions for any listing. Sellers can also run a title and bullets through ChatGPT, or write the questions out manually based on what real customers ask in reviews and support emails.
Traditional Amazon SEO Still Drives Most Sales [12:03]
The SEO pillar still drives most sales on Amazon. David's team targets over 80 keywords on a single listing, with each one appearing on average three times across the title, bullet points, back-end and Premium A+ Content. Top keywords appear 35 to 40 times.
And before anyone panics about keyword stuffing — David's continuous testing shows the conversion penalty is negligible (one or two percent) compared to the traffic gain (ten to twenty percent or more). "Keyword stuffing" is a fear from the Google 2012 era. On Amazon in 2026, frequency is still a seller's friend.
Why Now, Not Next Year [19:30]
David is urgent about Rufus optimisation, and there's a specific reason. Amazon SEO used to be like this. In 2016 to 2018, sellers were building empires on a single product image and a short title because almost no one was doing proper SEO. The brands that put in the work then are still ranking today, a decade later, on the authority they built when the field was empty.
Rufus is in that 2016 moment. Adoption is on a hockey-stick curve. The questions a seller teaches Rufus to associate with their product now will compound into 12-24 months of defended position. The brands that wait will be fighting for second place, the same way most SEO sellers are today.
Why Not A/B Testing Your Main Image Is "Criminal" [24:57]
Once a product is showing up — in Rufus answers and on search pages — people still need to click. And 80% of click-through rate is driven by one thing. The main image.
David did not soften this. "Not doing at least two or three A/B tests for a product that's doing over $1,000 per month in sales is criminal in 2026."
The maths is brutal in a good way. A 10% lift in clicks is not a 10% lift in sales — it's much bigger over 12-24 months, because more clicks tells the Amazon algorithm a product is performing better than the competition, so Amazon shows it more. The compounding can be the difference between $40,000 and $50,000 in monthly revenue from the same listing.
A few practical rules from David —
Make the product as large as possible in the frame
Add contrast, especially heavy shadows behind white products that disappear into a white background
Test the image at the size of two thumbs on a phone, because that's the actual viewing surface for most shoppers
Add a label on or attached to the product showing the main keyword (a "blue football" label on a blue football). Even a 5% subconscious lift in click-through is significant
Don't break the label free from the product — Amazon's AI flags floating elements

Conversion — Where Premium Branding Earns The Price Tag [33:08]
The conversion pillar is where most sellers leave the most on the table, because it's the largest surface area to work with. Seven product images. Premium A+ Content. The brand story, a section above or below the product description that almost everyone skips.
David's principle for product images is the same as for Rufus listings — answer the top 5 to 10 questions in priority order, but visually. If a seller is selling a £60 ($75) premium product against £12 ($15) competitors, the first image after the main one needs to be a "us versus them" answer. As David put it, "a confused buyer is not a buyer."
One of David's clients, Doodlebrush, sells at £60 ($75) against competitors at £12 ($15) and holds 20% market share. The reason is branding. Premium pricing demands premium branding — the same logic that lets Nike and Apple charge what they do. If the branding doesn't match the price tag, the price tag loses every time.
Text For The Algorithm, Images For The Human [39:32]
David's closing line is the one Matt said he'd be quoting. "The text speaks to the algorithm. The images speak to the customer."
The titles, bullet points and back-end fields aren't there for the customer to read (almost no one does). They're there for Rufus and the search algorithm to chew through and decide where to rank a product. So write that text for the machine — keyword-rich, question-answering, dense.
The product images are where the human meets the product. They need to be visual, scannable, and built around the questions a real buyer is asking — not paragraphs of text the algorithm could already read elsewhere.
Where To Start This Week
Audit the top 25 questions — through Selluna's free tool, ChatGPT, or manual research from reviews
Rebuild the title and bullets to answer those questions in priority order
A/B test the main image on any product doing $1,000+ per month with Brand Registry. Run for at least four weeks
Audit the conversion stack — seven images, Premium A+ Content, brand story. Each one needs a job
Match the branding to the price tag — premium pricing demands premium branding before customers will accept it

Today's Guest
Today's guest: David Balan
Company: Selluna
Website: https://www.selluna.ai/
LinkedIn: Connect with Selluna on LinkedIn
Episode link: Up to 40% of Amazon US sales now touch Rufus, Amazon's AI shopping assistant — and most sellers haven't optimised a single product for it. That's the gap David Balan, founder of Selluna, walked through with Matt this week. The customer who used to type "blue football" into a search bar is now asking a chatbot what kind of football would suit a six-year-old, and the products that show up first are the ones answering the right 25...

Rachel Hanretty pays £500 a month for a tool that fact-checks her marketing agency — and she has opinions about why every seasonal ecommerce founder should do the same.
In this episode, Matt Edmundson sits down with Rachel Hanretty, founder of Mademoiselle Macaron, the Scottish brand shipping up to 25,000 macarons a week UK-wide. Rachel built the business from a St Andrews student flat in 2013 after training in Paris, and thirteen years on she has strong views on attribution, agency accountability, the seasonality trap, and the gulf between what Instagram wants and what authenticity actually looks like. The conversation covers Triple Whale, Klaviyo's tendency to overstate its own impact, a Pinterest wedding ads experiment that burned thousands, a Charlotte Tilbury collaboration that produced lipstick-matched macarons at H Beauty, and why Rachel feels "gaslit by her own business" every April to September.
It is half ecommerce masterclass, half therapy session — and Rachel is refreshingly honest about both.
In This Episode
(01:30) From a Paris kitchen to 25,000 macarons a week
(07:15) The seasonality trap: "gaslit by my own business"
(14:00) Attribution wars — why Klaviyo and agencies overclaim
(17:00) The £500/month tool that fact-checks her agency
(22:00) Everyone needs an Alan: fractional CFOs and complementary hires
(27:00) ROMS — return on macarons spent, and experiential retail
(34:00) How do you do authentic Instagram when your niche is Paris?

Attribution Is Broken, And Every Platform Is Overclaiming
(Around 14:00)
Rachel's biggest challenge is not product or operations — it is knowing which marketing pound is actually doing the work. Klaviyo, she says, is "way too generous" with the revenue it claims. Agencies do the same. Until she started using Triple Whale, she had no independent way to push back.
"When you're being charged thousands of pounds for a service, you're like, no, I'm really sorry, but that number doesn't match up with the number in my bank." — Rachel HanrettyMatt picks up the thread with a story from Neil Hoyne, Google's Chief Measurement Strategist and author of Converted, who once traced a single pair of shoes back through 236 separate customer touchpoints. If every platform claims that sale, the maths does not work. Matt's recommendation (crediting previous EP guest Matt Putra): stop fighting about platform-level ROAS and track blended ROAS — the one number no agency can spin.
The £500 Tool That Coaches Her Through The Data
(Around 17:00)
Rachel pays £500 a month for Triple Whale, but the attribution modelling is not actually what she values most. The real prize is the platform's AI chat — she calls it "MobiChat" — which she uses as a coaching tool.
"I just say, can you explain what's going on with the Meta ROAS? Should I cut the budget? And it throws up the halo effect model of Meta on Google. I'm like, thank you so much for explaining it — because you're not paying by the hour." — Rachel HanrettyOver six to nine months she has used the tool as much to build her own knowledge as to interrogate the data. When an agency sends a report, she pastes it into Triple Whale and asks: is this right? That single habit has changed the power dynamic in every agency conversation she has.
Shopify-only tool (a limitation worth flagging for non-Shopify founders)
Value comes from the coaching layer, not just the attribution model
Used to sanity-check agency claims and reduce reliance on fortnightly calls

"Gaslit By My Own Business" — The Seasonality Trap
(Around 07:15)
October to March is the heroic half of Rachel's year: Halloween, Christmas, Valentine's Day, Mother's Day, Easter. April to September is a different business entirely — different audience, different messaging, different paid media strategy. The same product, but the customer and the context shift underneath her.
"Sometimes I feel like I'm being gaslit by my own business." — Rachel HanrettyHer Pinterest wedding-ads experiment is a case in point. The team put thousands into Pinterest last summer aiming at brides, only to realise that people already getting married that summer had booked their suppliers months earlier. Wedding customers need to be reached a year ahead — not in the season itself.
Rachel's lesson: cost per acquisition, not ROAS, is the number that matters when you are used to seeing 5-8x returns and suddenly cannot.
Every Company Needs An Alan
(Around 22:00)
Rachel started the business at 23 with very little reference point for managing people. Two leadership coaches and one fractional CFO later, she has a working philosophy: hire for the gaps in your own skill set, and do not be afraid when those people outshine you in their area.
Her fractional CFO is called Alan. He builds the dashboards that link a website order to grams of sugar, the cost of a 20% off promo, and the capacity implications of a given campaign. He is, as Rachel puts it, "the adult in the room."
"Every company needs an Alan." — Rachel HanrettyMatt echoes the point with Andrew Carnegie's epitaph: "Here lies a man who knew how to surround himself with people cleverer than himself." The discipline is not just hiring well — it is managing your ego when someone else shines.
ROMS, Charlotte Tilbury, And The Rise Of Experiential Retail
(Around 27:00)
Rachel's framework for the business breaks down into three "moments of luxury": moments of celebration (weddings, birthdays), moments to impress (corporate gifting and brand activations), and moments of thoughtfulness (gifting). The middle category is where the most interesting growth is sitting.
Last year Mademoiselle Macaron worked with Charlotte Tilbury across H Beauty stores, producing thousands of macaron boxes colour-matched to individual lipsticks. The end customer may or may not eat the macarons — they are often photographed and posted on social media alongside the product. Rachel coined a term for measuring this kind of spend: ROMS — Return on Macarons Spent.
"Sometimes I don't even know if the macarons get eaten, frankly. I'm okay with it as long as they were paid for." — Rachel HanrettyThe Freezer Confession (And A Note From Pierre Hermé)
Rachel shared an operational secret that many founders in the fresh-food space will recognise: Mademoiselle Macaron freezes its macarons. Far from being a downgrade, the French pastry chef Pierre Hermé has publicly said that "anyone who says they don't freeze their macarons is a liar — because it improves the texture." Freezing makes the texture more marshmallow-like, and it gives Rachel the operational flexibility to take pre-orders from October and ship fresh-feeling product right up to the last Christmas posting day.
Authenticity vs The Algorithm
(Around 34:00)
Rachel's "question for Matt" was the most honest moment of the conversation: her niche is Paris, which is aesthetically beautiful, so every other bakery brand on Instagram is polished to within an inch of its life. She went to Paris in January, paid for an Airbnb-experienced photographer to shadow her for 36 hours, and hated every minute of it. Her Scottish phrase for how she feels about polished-female-bakery-Instagram: "gie's me the boak" — it makes her feel ill.
Matt's advice: stop trying to be polished. The videos doing one to two million views are shot on iPhones. The accounts winning are the raw, authentic ones. He points to his own son, Zak the Nutritionist, who built 70,000 Instagram followers in six months doing 60-second IBS-help clips filmed in Matt's back garden — and has now turned that audience into a successful recipe book.
"Consistency wins every single day and authenticity wins every single day. It does not have to be polished. People buy you." — Matt EdmundsonKey Takeaways
Fact-check your agencies. Platform-level ROAS from Klaviyo, Meta, and Google are all overstated. Use blended ROAS and an independent tool to sanity-check claims.
Cost per acquisition > ROAS when you move between seasons or audiences.
Long-tail channels need long-tail thinking. Pinterest for weddings works — but the year before, not during the season.
Hire your Alans. A fractional CFO (or anyone filling your skills gap) is worth more than another marketing hire when you cannot read your own numbers.
Experiential retail is a real revenue channel. Brand activation work pays per unit and builds social proof.
Stop trying to be polished. The algorithm rewards authenticity, not aesthetic perfection — even in aesthetically-driven niches.

Resources Mentioned
Mademoiselle Macaron —

Most ecommerce sellers limit themselves to one or two marketplaces. Jorrit Steinz, CEO and founder of ChannelEngine, argues that's leaving serious money on the table. With over 1,300 marketplaces now available worldwide, the opportunity to reach new customers has never been bigger — or more manageable with the right tools.
In this episode of The eCommerce Podcast, Matt Edmundson sits down with Jorrit to discuss why marketplace selling is no longer optional, how to choose the right platforms for your products, and why spending on marketplace ads might be a smarter move than pouring budget into Google. Jorrit also shares practical advice on getting started without drowning in manual processes, and reveals the biggest mistake most sellers are making right now.
Key points from this episode:
01:49 — What ChannelEngine does and the 1,300-marketplace landscape
09:17 — How to choose which marketplaces to sell on
20:47 — The biggest mistake in marketplace selling
25:49 — Why marketplace ads beat Google spend for organic ranking

The 1,300-Marketplace Opportunity You're Probably Missing
[01:49] Jorrit explains that ChannelEngine connects brands and retailers to over 1,300 marketplaces globally, handling product listing, order management, and stock synchronisation in one platform. The landscape has shifted dramatically. COVID accelerated the trend of traditional retailers launching their own marketplaces, and Amazon itself is pushing sellers from 1P vendor relationships to 3P seller accounts — particularly those under a $20 million threshold.
Matt shares his own experience with this shift, noting that Next in the UK now operates a marketplace where brands sell alongside Next's own range. These "local hero" and vertical specialist marketplaces are multiplying fast.
"There are 1,300 marketplaces in our network. Local heroes, vertical specialists — the landscape is much bigger than most people think."How to Pick the Right Marketplaces for Your Products
[09:17] Jorrit's advice for choosing where to sell starts with a simple principle — look at the seller-to-buyer ratio on each marketplace, not just total traffic. Amazon has 600 million products listed, meaning competition is fierce. Smaller, more specialised marketplaces often have far fewer sellers competing for the same buyers.
He recommends using AI tools like Claude or Perplexity to research marketplace options in your category. The practical starting point is modest — six to seven marketplaces in the first year, then build from there. One of his German customers now runs 60 marketplaces with just two people, because the right automation makes scaling manageable.
"Look at the seller-to-buyer ratio on each marketplace. Amazon has 600 million products — competition is fierce."Why the Biggest Mistake Is Not Selling on Marketplaces at All
[20:47] When Matt asks about common marketplace mistakes, Jorrit's answer is direct — the biggest mistake is not doing it at all. Many brands hesitate, worried about channel conflict or complexity, and miss the growing opportunity cost entirely.
For established brands, Jorrit recommends pairing an agency with automation tools to get moving quickly. For those just starting out, the path is different — begin on marketplaces first, then build your direct-to-consumer shop alongside. The key is to always be where your customers already are, rather than trying to pull them to where you want them to be.
"The biggest mistake? Not doing marketplace selling at all. The cost of not doing it is increasing every day."Spend on Marketplace Ads, Not Google — and Why That Builds Organic Ranking
[25:49] Jorrit shares a contrarian advertising strategy. Rather than spending on Google Ads to drive traffic to your own store, invest that budget in marketplace advertising. The reason is that marketplace ad spend builds your organic ranking within that platform. As your products climb the rankings, you get sustained visibility long after the campaign ends.
He also recommends driving external traffic — from social media, email, or other channels — directly to your marketplace listings rather than your own website. This signals demand to the marketplace algorithm and pushes your products higher in search results. It's a compounding effect that Google Ads simply cannot replicate.
"Spend on marketplace ads, not Google. It builds your organic ranking on the platform."Today's Guest
Today's guest: Jorrit Steinz
Company: ChannelEngine
Website: channelengine.com
LinkedIn: Connect with Jorrit on LinkedIn
Episode link: https://www.ecommerce-podcast.com/why-your-products-should-be-on-60-marketplaces-not-just-amazon

Start a Print on Demand Business on Etsy for Less Than £15
What if you could launch a profitable ecommerce business with no stock, no warehouse, and no design skills for under £15? Nigel Wymer from POD Launch Pro joins Matt Edmundson to walk through exactly how to build a print-on-demand business on Etsy from scratch. With nearly 20 years in ecommerce and almost £40k net earned last year while barely working on his stores, Nigel brings the receipts to match the strategy. In a hands-on episode framed around helping Matt's daughter Zoe start her first business, they cover niche research with AI, Etsy store setup, the algorithm game, sourcing, pricing, and realistic income expectations.
Key points from this episode
[04:32] What print on demand actually is and why it requires zero stock
[07:28] Why Etsy beats Shopify for beginners
[12:41] How to research niches using AI and why specificity wins
[33:33] Setting up your Etsy store for SEO and trust
[36:41] Working with the Etsy algorithm as a new seller
[45:59] Pricing without racing to the bottom

Why Etsy Beats Shopify When You're Starting Out
[07:28]
Platform choice is one of the first decisions any new print-on-demand seller faces, and Nigel's advice is unambiguous. Start on Etsy.
The reasoning is simple. With Shopify, you launch to zero traffic. From day one, you're spending on ads or pouring time into social media just to get eyes on your products. Etsy, by contrast, had 86.6 million buyers last year — people who actually purchased at least one item. That's a massive built-in audience you can tap into without spending a penny on advertising.
"If you want to go on to Shopify, then that's something I recommend you go on to later, once you know what sells. You've got the audience there now that you can trial it with on Etsy." — Nigel Wymer [08:10]The trade-off is familiar to anyone who's sold on Amazon. You don't own the customer relationship the way you would on your own site. But for someone just starting out, access to traffic far outweighs that limitation. And with just a £15 one-off fee to open a store, the financial barrier is almost non-existent.
Niche Research With AI Changes Everything
[12:41]
This is where most people get it wrong. They dream up a design, create a product, list it, and hope for the best. Nigel's approach starts in the opposite direction — research first, always.
And there's a common misunderstanding about niches. Dogs and cats are not niches. Sports is not a niche. You need to go much deeper.
"You really want to go into these smaller areas where you are being very specific in your targeting. But also there is much less competition." — Nigel Wymer [14:40]Nigel demonstrated this live on the show. Starting with "sports" as a broad category, he used AI to identify underserved sub-niches. Ice hockey came up. Still too broad. So they went deeper — ice hockey goalkeepers, then into the specific slang, culture, and identity of that community. That's where you find products so targeted that customers feel like you made them just for their world.
The obvious pushback is market size. Nigel's answer is elegant — you stack niches. Each micro-niche might only generate a small amount on its own, but when you're targeting dozens of specific communities, it adds up significantly over a year.
"What you do is you're stacking up lots of different niches... you're getting a small amount of money from each, but stacking it up over the year will add up to a lot of money." — Nigel Wymer [15:58]Matt ran a live AI experiment during the episode, asking Claude for 20 underserved print-on-demand niche ideas for Etsy. The results included adaptive clothing for neurodivergent communities, grief and memorial keepsakes, and niche outdoor micro hobbies like gravel biking and disc golf.
Working the Etsy Algorithm as a New Seller
[36:41]
New stores face an inevitable chicken-and-egg problem. No reviews means less trust. Less trust means fewer sales. Fewer sales means no reviews.
Nigel's solution requires patience rather than tricks. List regularly — not ten products in one day, but one today, another tomorrow. Each new listing gets a temporary visibility boost from the Etsy algorithm as it tests whether anyone is interested. By listing consistently, you're telling the algorithm you're an active seller, which earns higher rankings.
"Once you've got 10 sales, then the algorithm knows what you're about and who to show your product to. And it will then start growing." — Nigel Wymer [37:50]After that initial growth phase, you don't need to maintain the same pace. Nigel spent most of last year barely adding new products because he was renovating a house, and his established stores kept generating sales on autopilot. The key elements for a professional store that builds trust from the start include a branded banner (made on Canva), a simple logo, and a photo of yourself — since over 90% of Etsy sellers are individuals, and buyers want to know who they're purchasing from.
The About Me section does double duty. It builds trust with buyers and contributes to your store's SEO. Beyond that, it's standard ecommerce optimisation — keyword-rich titles, detailed descriptions, proper meta text, tags using keyword phrases rather than single words, and alt text on images.
Realistic Income and Getting Started This Week
[48:23]
Nigel cleared almost £38,000 net last year while barely working on his stores. That's the established-store reality. For someone just starting out, the benchmarks are different but still compelling.
Nigel appeared on a Channel 4 programme presented by Scarlett Moffatt that investigated different side hustles. Print on demand came out on top as the easiest way to make £10,000 a year.
Pricing is straightforward and doesn't require competing on price. A basic mug costs Nigel around £4.50 plus shipping from the manufacturer. He sells it for £10.99 plus shipping, passing the shipping cost through to the customer. For international orders, larger manufacturers have production facilities in multiple countries, so a US order gets made in the US — keeping shipping costs and delivery times reasonable.
"The answer is not, as a lot of people will do, I'm not getting sales, I'll discount it, I'll go cheaper than everyone else. That's not the answer. The answer is create your totally unique product." — Nigel Wymer [45:59]To get started this week, open an AI tool and ask for 20 underserved print-on-demand niche ideas for Etsy. Pick one niche and go deeper — ask for sub-niches, community slang, inside jokes, and cultural references. Set up a free Etsy store for £15, create a banner and logo on Canva, write an About Me section, find a UK-based print-on-demand manufacturer, and create 4-5 products to list gradually over the first week. Then keep listing regularly.
Today's Guest
Today's guest: Nigel Wymer
Company: POD Launch Pro
Website: podlaunchpro.com
LinkedIn: Connect with Nigel on LinkedIn
Episode link: https://www.ecommerce-podcast.com/how-to-start-a-print-on-demand-business-on-etsy-with-nigel-wymer

Most ecommerce operators can rattle off their LTV, churn rate, and CAC without thinking. But when Jay Myers asked a room full of subscription experts at SubSummit whether anyone knew their referral rate, only one person raised their hand — and his was zero. That single overlooked metric, Jay argues, is the difference between a business that flatlines and one that grows exponentially.
In this episode of The eCommerce Podcast, host Matt Edmundson sits down with Jay Myers, co-founder of Bold Commerce, to unpack the data behind referral-driven growth — and why a tiny shift in referral rate can mean 56 times more subscribers. Jay shares the strategies that actually work, from golden ticket referrals to paid memberships, and explains why most "share and save" programmes fall flat. Whether you run a subscription box or a one-time-purchase store, this conversation is packed with ideas you can start using today.
00:00 — Introduction and Jay's background
02:56 — From archery store to 32 Shopify apps
11:14 — The launch of RePete, Bold's AI reorder agent
14:32 — The Subscription Death Curve and the metric nobody tracks
22:48 — Golden tickets and the psychology of scarcity referrals
36:40 — Practical implementation for any ecommerce brand
40:14 — Dollar credits vs discounts vs percentages
45:45 — Paid membership vs free loyalty programmes
49:21 — How to connect with Jay

Why Your Referral Rate Matters More Than LTV, Churn or CAC
[14:32] Jay's 2022 SubSummit talk, "The Subscription Death Curve," revealed a striking blind spot in ecommerce. He asked a room of subscription operators to name their most important metric. People shouted out LTV, churn, CAC, and ARPU. Then he asked how many knew their referral rate. Not a single hand went up.
The data tells a compelling story. Doubling acquisition moves the subscriber count higher, but the business still flatlines at the same point — around month 32. Reducing churn pushes the flatline further out, but it still flatlines eventually. However, increasing the referral rate from just 0.8 to 1.2 — a difference of 0.4 — results in 56 times more subscribers by month 30.
Less than 3% of the average marketing budget goes toward referral programmes
A viral coefficient above 1 is needed for true viral growth
Super-referrers need uncapped ability to keep sharing

"There is no bigger thing you can do to your business to change the long-term health of it than increasing customers who refer other customers." — Jay MyersGolden Tickets and the Psychology Behind Referrals That Actually Convert
[22:48] Most referral programmes follow the same formula — share a code, both people get a discount. Jay explains why this generic approach rarely works, and it comes down to the difference between intrinsic and extrinsic motivation.
The golden ticket strategy works on a completely different principle. Instead of rewarding the sharer upfront, the approach relies on scarcity. One unique code, no immediate reward for the person sharing, and a focus on finding the ideal customer rather than casting a wide net.
Generic "share 10, get 10" codes produce low engagement because extrinsic rewards crowd out genuine enthusiasm
Scarcity-based referrals tap into intrinsic motivation — people share because the offer feels special
This strategy requires no app — Jay walks through how to run it with a CSV export and any email tool

Dollar Credits Beat Discounts — and Here's Why
[40:14] Not all incentives are created equal. Jay breaks down the psychology behind different reward structures.
Dollar credit performs best — it creates a sense of ownership ("there's £10 in your account")
Dollar discount comes second — tangible but framed as future savings rather than something already owned
Percentage discount performs worst — abstract, requires mental arithmetic, feels distant

The distinction is subtle but significant. Credits feel like money already in your pocket, which changes buying behaviour far more effectively than a promise of future savings.
Paid Memberships and the Sunk Cost Psychology That Builds Loyalty
[45:45] Jay highlights three brands that show why paid membership programmes outperform free loyalty schemes through sunk cost psychology.
Restoration Hardware — $125/year membership. Their 100,000 members account for 95% of revenue and spend 400% more than non-members
Fabletics — $50/month gives members $70 in credit. The monthly commitment fundamentally changes how members shop
Amazon Prime — almost never the cheapest option, but Prime members rarely price-shop elsewhere. The membership creates brand affinity that makes comparison redundant

When someone pays to be part of something, they behave differently. Free loyalty programmes ask nothing and, in return, tend to get very little engagement. A paid membership creates a psychological contract that drives repeat purchases and deeper brand connection.
Today's Guest
Today's guest: Jay Myers
Company: Bold Commerce
Website: boldcommerce.com
LinkedIn: Connect with Jay on LinkedIn
Episode link: https://www.ecommerce-podcast.com/the-metric-nobody-tracks-that-drives-56x-subscriber-growth