24 months. 2.2M Shopify orders. $274M. AI drove 0.1% of it.
"This traffic is nothing. It is a rounding error."
Kurt and Paul pulled 24 months of raw order data from 10 established Shopify stores: 94 million sessions, 2.2 million orders, $274 million in revenue. They went looking for the AI shopping revolution and found 0.1% of revenue. The catch: those referrals convert better than any other channel, and getting ready for them takes one afternoon of free work.
Then the discount data. Across $285 million in gross sales, nearly every store hands the same 20% code to new and returning customers alike, all year long, while the promo type that wins on every metric gets 7% of the budget.
Also covered: the channel quietly outselling all AI sources combined 35 to 1, and the leaked evergreen code that cost one store $349,000.
Swym - Wishlists, Back in Stock alerts, & more
getswym.com/kurt
Cleverific - Smart order editing for Shopify
cleverific.com
Zipify - Build high-converting sales funnels
zipify.com/KURT
Apply for Shopify Help
ethercycle.com/apply
See Our Results
ethercycle.com/work
Free Newsletter
kurtelster.com
The Unofficial Shopify Podcast is hosted by Kurt Elster and explores the stories behind successful Shopify stores. Get actionable insights, practical strategies, and proven tactics from entrepreneurs who've built thriving ecommerce businesses.
Speaker A:
This episode is brought to you in part by Swim. Here's the thing about wishlist apps. Most of them just sit there. A customer saves a product and then nothing happens. Swim actually activates that data. When someone wishlists a product, you can trigger price drop or back in stock alerts and feed that intent directly into Klaviyo or your CRM. You're not guessing what people want because they've told you. Plus, customers can share wishlists for gifts and your team can view them to offer personalized service online or in-store. And unlike cart abandonment, wishlist data is permission-based. These are people raising their hand saying, "Hey, I want this, just not right now." Swim's been around for over a decade. It powers 45,000 stores and installs in about 5 minutes. You could try it for free today at getswim.com/kurt. That's G-E-T-S-W-Y-M.com/kurt. So, you know, it's not unlike me. I'm an oily monster.
Speaker B:
I'm like, what? I'm the one that leaves grease stains on the roof of my car.
Speaker A:
That's a little funny. I enjoyed that. What is this weird stain on the roof of my car? I'm like, ah, it's leaky sunroof. No.
Speaker B:
That's my head grease. Yeah.
Speaker A:
Just a very tall man.
Speaker B:
That's my head grease.
Speaker A:
And a light colored roof in a car, not a good headliner. But all right, on our last episode where you and I recorded a couple weeks ago, we're like, hey, how much traffic in a high volume store, so 8 figures, How much of that was generated by AI?
Speaker B:
And that's $1 trillion? Is 8 figures?
Speaker A:
Yeah. No, it's $20 million.
Speaker B:
Oh, okay.
Speaker A:
Greater than $10 million. Yeah. Less than $10, but more than $1 million, 7 figures.
Speaker B:
I love that movie about the Black ladies in the space program.
Speaker A:
Numbers?
Speaker B:
That's Hidden Figures.
Speaker A:
Hidden Figures? I was like, what's the name of that movie? No, okay. I was like, guess what it is?
Speaker B:
Oh yeah, it was like under $5,000. Yeah. They got no, there's no AI-directed revenue coming into that giant store. There we go. That's what we talked about. Now I remember.
Speaker A:
So, all right. Everybody, everyone says they're data-driven, right?
Speaker B:
Yeah. Until they don't like the data. Then they're like, yeah, that's not true for various reasons. Even though you might think that that's true, but that data's wrong.
Speaker A:
I'm data-driven. I make data-driven decisions until it doesn't agree with me. But all right, we have data, 24 months of order data analyzed out of 10 stores. So is this statistically significant? Probably not. Is it anecdotal and interesting? Yes.
Speaker B:
Well, and here's the first caveat I saw was that your time period was way too long. Because no one used AI 2 years ago.
Speaker A:
Okay.
Speaker B:
So there were all those months where there was no AI traffic that doesn't count in.
Speaker A:
Yeah, well, we'll get into the objections, but the short answer is 24 months of raw data, 10 established Shopify stores across a variety of verticals. So, you know, some good variety here. It's not all concentrated in one area and 94 million sessions. 2.2 million orders. We're looking at data from July 2024 through 2026. $274 million in revenue analyzed here. So of our $274 million in revenue over 2 years, what percent of that is referrals from AI? So ChatGPT is number 1 by far. And then Bing, Perplexity, et cetera.
Speaker B:
It's on the sheet in front of me. So I could just say it's 0.06%.
Speaker A:
That's right.
Speaker B:
Over the 2 years.
Speaker A:
Okay, but the pushback was, well—
Speaker B:
That's not 6%. You know how 6% would be like 0.06 if you're doing the math? It's 0.06%.
Speaker A:
Yes.
Speaker B:
Less than 1%.
Speaker A:
Less than 1/10 of 1%. That's if we go over 2 years. But fine, 2 years ago, maybe normies weren't using AI. Certainly we were. But apparently in this cohort, that's the pushback is no. So let's look at year to date. What does it give us? What's it jump to? It hugely increases if we look at year to date. What is it?
Speaker B:
2.1%. So 1/10 of 1%.
Speaker A:
Yes. Okay. So technically it's increasing, but not by a ton. There was, all right, what the other pushback I got was, well, they asked AI and then they just Googled it.
Speaker B:
I ask AI, the AI gives me an answer and a link probably to the store, but I didn't click on that link. I actually then took the knowledge of that product and Googled that product and went through Google and then clicked on that link.
Speaker A:
Yeah. And so that's what we're actually seeing.
Speaker B:
So there's actually a ton of secret underground AI referrals going on. You just can't see 'em.
Speaker A:
Just can't see it. But all right, so ChatGPT went from 59 referrals in our dataset to 5,400 total. So, all right, it does climb, but it's not great. You know, our 5,400 referrals still puts us at 0.1% of revenue. So we really don't have— if that is true, it's like, all right, I ask what are the best socks and it's going to do the research and then give me, tell me The idea is I don't click on anything there. I then turn around and start a new session in a browser and look at it there by then Googling the brand name. This is the claim. If that were true, I really would expect to see that in like Google search and direct traffic. Yeah, you don't.
Speaker B:
I guess. Yeah, maybe there's just—
Speaker A:
I can't find a single piece of evidence to support it.
Speaker B:
I mean, so what we're saying is, is that, okay, it grew. It's the old disco stew where it grew. 6x year over year, but it went from, you know, 0.001% to 0.1% or whatever, 0.015, whatever 1/6 of 1/10 of 1% is.
Speaker A:
Yeah, we went from 0.01 to 0.6.
Speaker B:
So even if it's 6x again, it'll be 0.6%. And then maybe it'll 6x again. And so then that will be 3%.
Speaker A:
Yeah. I use Claude daily. I'm a power user. I'm using it for software development. I'm using it entirely in IT context.
Speaker B:
I have to hear you talking about this with other people who are also utterly brain damaged by AI and nonstop using it. And they're like, I set up an MCP server.
Speaker A:
I prefer Claude pilled to brain damaged.
Speaker B:
I set up an MCP server for my Opal Claw thing and it's using MD file. Absolute top 0.01% shit. Whereas all the vast majority of users are like, what should I eat for lunch? But you guys are talking and you're like, hmm, yes, this is how the people are using AI. I would know. I have insights on that. It's like you're going to the Sorbonne in Paris and you're just like, how do I like, how do you think people make burgers? And then they're like, well, clearly here's how you would make a burger. It's like they don't know how people make burgers. Like you're so completely beyond the burger making world that you have no concept of it anymore. That's you trying to talk about normal people using AI.
Speaker A:
100%. I think, you know, if your exposure to AI is X, is the Twittersphere, you know, on there it very, you know, in a tech circle, very much feels like unless you are on the bleeding edge of this, you are falling behind. But then the reality of like actual consumer adoption, most of it is I have a free ChatGPT account I use occasionally. Yeah. And I always say, like when referring to average person's usage, I say ChatGPT, not agent, you know, not AI, LLM, whatever, because in these referrals, it is almost exclusively ChatGPT. Like that's when you say AI to a normal person, you may as well just be saying ChatGPT. They're synonymous at this point.
Speaker B:
So what we're trying to say here is this traffic is nothing.
Speaker A:
Yes.
Speaker B:
It is nonexistent. It is a rounding error. It is a rounding error. By far, it is a rounding error.
Speaker A:
It's a slowly, slowly increasing rounding error.
Speaker B:
Well, it's not, I wouldn't say it's slowly increasing, but when you're increasing from zero, it's gonna take a long time till it's anything.
Speaker A:
Yeah.
Speaker B:
And what we learned in these insights, this traffic is better. It converts better than—
Speaker A:
It's majority new customers.
Speaker B:
It's new customers. It converts better, all this other stuff. Um, it seems like the real traffic you would want, but I don't think it's an adequately sized cohort to claim that. Mm, is my claim, is my motivated reasoning here, because I don't like this data. See, I'm very data-driven as well, and I don't like this data. So here's why it's bad.
Speaker A:
Explain it to me.
Speaker B:
Is this is such a tiny amount of data. Again, 5,400 referrals a month. That amount of traffic, you can't tease anything out of that shit. And so this idea that this traffic all converts better, it's a majority new customers, like these ads are 25% new customers and the ChatGPT stuff is 30% new customers. It's like, what are the error bars on this utterly minuscule amount of traffic? You don't know.
Speaker A:
Well, that's what's frustrating. I have over 2 million orders.
Speaker B:
Yeah.
Speaker A:
I have a huge set of data that I'm looking at. But within it, confirmed AI referrals are basically nonexistent.
Speaker B:
Yeah.
Speaker A:
And so then I'm limited to like my out of 2.2 million orders, it's less than 6,000.
Speaker B:
Yeah.
Speaker A:
Actually were AI referral. And of those, it's almost entirely ChatGPT. Okay. So within that though, I could figure out what's going on with it. Part of it is ChatGPT has a real tendency to just drop you on the product page. I think that helps. Oh yeah. And it is pre-selling you and you're really only going to see referrals from people who are interested. Like I'm not going to ask it, hey, tell me about socks. And then it's like, well, here's a link to a shirt.
Speaker B:
That's the thing. Well, in your Facebook ads, you're kind of cold telling them about socks.
Speaker A:
Yeah, I'm interrupting somebody scrolling.
Speaker B:
Yeah. Whereas ChatGPT, these people are already in the sock mindset of looking for socks. Yeah.
Speaker A:
So what's fun about this is like, okay, there is something here. Assuming ChatGPT could get out of their own way and, you know, really like adopt these shopping recommendations more consistently, be sending people, you know, to these stores. But then it comes down to, all right, who is ultimately using it? How many people are using it? Are they using it for shopping or are they just doing what they always did and they Google it? Or it's impulse purchases, or I need an influencer to tell me. And currently it's like, yeah, it just doesn't look like it's a ton of people.
Speaker B:
I can't argue with the quantitative numbers. It is what it is. That's the number of referrals we're getting. I poo-poo the qualitative numbers, the quality of this traffic, because I think it's too small of an amount to actually accurately determine whether this traffic is quality or not. But whatever, that is what it is. My end question here is, what should the store owner do about this knowledge? We have this report online. It goes through all of this.
Speaker A:
Yes, all of this is published.
Speaker B:
You've been making a lot of posts about it. Go check out Kurt's feeds in various spots. I'm sure he'll link to it in the show notes to this report. You're dumping all this information on me.
Speaker A:
Yes.
Speaker B:
Great. What am I supposed to do with this?
Speaker A:
I think two things. One, settle down. As someone who is in what is functionally a CTO role, I have FOMO. If I am not on top of the bleeding edge of whatever the hot tech is, then somehow I am failing at my job.
Speaker B:
You are the king of FOMO.
Speaker A:
Oh yeah.
Speaker B:
Tell them about your Reddit account, Kurt.
Speaker A:
Oh man. Marketing automation fail. I got a little aggressive and I plugged in, I guess called Composeo, but a service that will let you plug anything into an AI. And I saw that it would plug in Reddit, a thing that AI cannot access normally on its own, or at least it has to go workarounds. And I said, all right, we got it working. It was giving me trouble. I'm like, test it, test all its functionality. And Claude said, hey, are you sure you want me to post a comment under your real account? I said, well, just post one and then I'll delete it. That act immediately got the account banned from Reddit. A 7-year-old, 20,000 karma Reddit account, insta-banned.
Speaker B:
Because you absolutely needed to optimize your Reddit usage.
Speaker A:
Okay, I didn't, but the way I have adopted and learned AI in this really aggressive way is by just sacrificing home projects to it. 3 computers in my house are exclusive and my route, my network, my route, all this stuff exclusively managed by Claude Code. And yeah, I have broken stuff, but it's like, okay, I'm fine with that. My unfortunately my Reddit account was a sacrificial lamb, so not my favorite thing, but all right. Number 1 is like when you see numbers like this now, you don't have to— the FOMO can go away because there's nothing here Yet, which brings me to part 2.
Speaker B:
Fear of missing out. You're not missing out on anything.
Speaker A:
Yeah. Like, okay, whatever your revenue is, increase it by 0.6%. That is what you're missing out on. We can quantify this number for you. Support for this episode comes from Zipify, makers of ShopRank. Consider this: AI may be hiding your Shopify store and you wouldn't even know it. Today, more than 800 people use large language models like ChatGPT and Gemini to shop for their products. And those AI engines are deciding who to recommend. The question is whether they're picking you or your competitor. The trouble that most Shopify stores have errors they may not be aware of. Missing alt text, weak descriptions, JSON schema that is simply wrong. AI sees those errors and moves on without a second thought. And there goes the sale. Now you could do something about it. ShopRank checks your product catalog against 17 critical factors that can keep you from appearing in AI search. But unlike other audit tools, which only tell you what's broken, ShopRank fixes it for you automatically. No tech skills required. ShopRank gives you your store's AI visibility score in under 60 seconds and then helps you improve it for free. You could see your AI score right now by visiting zipify.com/ai. That's zipify.com/ai. Do take a moment to do this today before AI gives away any more of your sales. Number 2, it is not difficult to prepare for this future. There are plenty of people that will sell you recurring services and packages to do it. I don't know that you should do that.
Speaker B:
I don't, you don't need to buy any of that crap.
Speaker A:
Yeah, I think that's, you're just gonna get ripped off.
Speaker B:
Yeah.
Speaker A:
However, Shopify gives you the tools and makes this easy. If you're on a new and up-to-date theme, you have all the structured data in there. Great, got that. So like your SEO meta structured data, Okay, yeah, are there some optimizations? Sure, maybe. But like the majority of it is there out of the box, done. That is a requirement to be in the theme store. It's gotta support the structured data. Then the other one is the category metafields. That's where, you know, ChatGPT and other services could query Shopify through the catalog API or the Universal Commerce protocol, and it comes back with products. If you want your stuff to show up in there, guess what? Shopify defaulted you into it. You're already in it., but a complete product listing, just as in SEO, is what it takes to appear, as well as that filling out the category metafield. And I've mentioned this on the show, like when we were talking to Vanessa Lee at Shopify, you have to fill out this category metafield, that will then trigger additional metafields related to it, so it's like you're selling a backpack, backpack is an option, and then it'll be like, okay, what's the material, what's the color, what are features? Fill all that stuff out. All of that is context. The success of an AI's understanding how well it can impress you with, you know, its pretend thinking is based on how good its context is. And so that is your— that is your way in. If you have 10 products, if you have 100 products, you could just do this on your own, sitting at your computer in an afternoon, you know, go to the coffee shop, have fun with it. Congrats. You are now optimized. Shopify does so much of this and is leading the way on it.
Speaker B:
Yeah, I think, and that's my main thing is that it's like, this is all gonna be handled on like a structured data level. It's going to be Shopify's job. I mean, think about—
Speaker A:
Yeah, they're developing the tools.
Speaker B:
Let's say there's a universe where AI shopping utterly explodes and it becomes one of the new main ways that people, it becomes a channel like Google search, it becomes a channel like Meta ads, and it gets to that level. You don't think Shopify is just gonna be like, all right, here's the tooling to make all of our stores work perfectly with that. That's Shopify's problem, that's not yours. They're just gonna be left behind and allow their entire library of customers and stores to be left by the wayside? Of course not.
Speaker A:
Yeah.
Speaker B:
So like, this isn't your job, this isn't your problem to worry about this. You just gotta make sure, the thing you have to worry about is what Kurt just talked about. Make sure all the data you have on your products is correctly entered on the Shopify Product page, that's it. Shopify is gonna take that data and handle it however they need to handle it, but you don't gotta worry about that.
Speaker A:
Yeah, but if you don't fill the stuff out, and especially category metafield and then the subfields that come with it, it has no impact, it makes no difference on the front end of your store. So whether you fill it out or not, you're not gonna feel it, but the effort to fill it out is what gets you in these AI listings potentially.
Speaker B:
Yeah.
Speaker A:
That's how ChatGPT, ChatGPT tells Shopify, I'm looking for X, Shopify then uses those category metafields to figure out what is the right product to come back with. Then it hands 10 to 30 to ChatGPT. ChatGPT then decides what it's going to show the customer.
Speaker B:
And the ChatGPT stuff is actually a side quest because the actual reason Shopify wants you to upload all of this data about your individual products is the Shop app. The Shop app.
Speaker A:
The Shop app benefits just as much from this.
Speaker B:
Blows AI shopping out of the water.
Speaker A:
Okay. I'm glad you brought this up. Yes. I also, I'd like to bring up how years ago we made fun of Shop App. Just 'cause it's— Just for its name.
Speaker B:
Its bad name.
Speaker A:
Yeah. But, you know, bold move on Shopify's part. And it came full circle when I was in Times Square last Thanksgiving and there's Shop App ads running. And I'm like, oh my gosh, it's Shop App. That's from Shop App. My kids were like, okay, whatever. You know, didn't care, didn't get it. But for me it was cool. But all right, in your Shopify dashboard, you go in the Agentic dashboard. It's under sales channel. It's not a sales channel, it's a dashboard.
Speaker B:
Yeah, yeah.
Speaker A:
Under sales channels, Agentic. Open that bad boy up and it'll tell you ChatGPT, Bing. I think I forgot the Bing one.
Speaker B:
Bing?
Speaker A:
What the hell? What's Microsoft's called? Copilot.
Speaker B:
Oh.
Speaker A:
Yeah.
Speaker B:
I was like, damn, they got Bing in there?
Speaker A:
Yeah, like it shows you who your 4 referral sources are. Let me load one up as long as we're here. Yeah, it shows you like, it goes ChatGPT and the store I'm looking at, sales last 30 days, $800. Meta, zero. Microsoft Copilot, zero. Shop, $14,000.
Speaker B:
Yeah. I mean, Shop is 35 times bigger than all AI shopping combined.
Speaker A:
Yes.
Speaker B:
So however much you're worried about like, oh, ChatGPT shopping, gotta be ready for AI shopping, you should be 35 times more worried about your shop app.
Speaker A:
Yeah, when was the last time you looked at your shop settings?
Speaker B:
Your shop settings and how—
Speaker A:
You look at it at all?
Speaker B:
Yeah, how your store presents in shop app and how it looks on there.
Speaker A:
100%.
Speaker B:
That 35 times more than you care about AI.
Speaker A:
That is the headline, it's like, shop is AI shopping. You got to use the thing.
Speaker B:
Yeah.
Speaker A:
If you're in setting that up, it's not difficult to set up. You're really setting it up once.
Speaker B:
It's the same thing we just said. Make sure all of the granular data about your products are filled out on that product listing. That's the most important thing. Because then the Shop app can direct people to the exact product they're looking for. Hopefully it's one of yours.
Speaker A:
What, man, what's the last time you looked at the shop app.
Speaker B:
I never looked at the shop app.
Speaker A:
Do you have it installed?
Speaker B:
I think I do have it installed, maybe for like tracking. That was their little— that was their perfect genius little scam, is that whenever you get— whenever you order something and it's like, well, you could track your package on the shop app. Well, obviously I need to track my package. What are you crazy?
Speaker A:
I'm searching. I searched Cyberpunk Hot Wheels, did in fact find it immediately, and it found it for way less than they were going for on eBay this week. So it 100% worked.
Speaker B:
Obviously my Shop app home screen is all Utz chips.
Speaker A:
Mm-hmm. I love Utz chips. Yeah. Not a semiconductor. He's talking about potato chips.
Speaker B:
Yeah. You know, the Utz, the famous German semiconductor manufacturer, Utz.
Speaker A:
Utz.
Speaker B:
Utz.
Speaker A:
But yeah, I mean, that my takeaway was, okay, AI traffic, the tiny tiny amount of it that there is. And really it's ChatGPT. ChatGPT referrals convert well. They're extraordinarily difficult to get. Maybe it's people Googling you and we never see it, but good luck proving it. And the real star is shop. And for all the same reasons, it's like, okay, if they open that, it's a marketplace. You open that up, you're there to shop.
Speaker B:
All right. You made a second, your first report, The AI shopping was a bang-up success. People told you you were a good boy. So immediately—
Speaker A:
I live for approval.
Speaker B:
That immediately, that meant you had to go do it again. So what was your second big boy, big data report?
Speaker A:
My other big boy, big data report was discounting.
Speaker B:
All right.
Speaker A:
And I have an agenda with this one.
Speaker B:
Well, you know, you know what we're going to tell you at the end of this.
Speaker A:
Yeah. Yeah. We have a free gift with purchase app called Promo Party. Yeah. Let's just disclose that now. Yeah.
Speaker B:
And the reason we're telling you, but This is all true, this is all real actual data. Free gifts outpace straight discounts by a ton in terms of both average order value, profit margins, on every metric, a free gift kills a discount, a straight discount.
Speaker A:
Yeah.
Speaker B:
In terms of what you get as the store owner.
Speaker A:
So same process, like I already had all that data pulled, now we could analyze it. And first I gotta get rid of all my free orders. I gotta get rid of my draft orders, B2B orders, all that goes away. I don't want my data muddied. So now I've got my clean set of order data. There's no customer info in it. I'm purely looking at line items. And what I discovered was the pitch on discounts is, well, you discount to acquire a customer, and then later on our returning customers, they won't get that discount. And we're gonna make up a customer lifetime value, the cost, the extraordinary acquisition cost of paying for my Meta ads or whatever it was. And it turns out everybody's just handing out the same amount of discounts to everyone always. Like if you just go, "Kurt, I have a Shopify store." All right, well, you probably discount 20 to 30%, almost definitely it's 20%. 100% you're discounting 20%, and it's gonna be on 35 to 40% of your orders, and whether it's a new or returning customer makes no difference. I think that's, you know, for as sophisticated as everybody on Twitter says they are, that this should not be the case.
Speaker B:
Yeah.
Speaker A:
The returning customers should not be getting that big discount, or maybe they should, but like, I just did not expect new and returning to be essentially the same.
Speaker B:
Yeah, they're just giving away significant percentage of their profits just to get the cheap hit of being like, see that, another sale. Look at that revenue by giving straight discounts to people. Yes.
Speaker A:
And then it's—
Speaker B:
Willy-nilly with no rhyme or reason.
Speaker A:
I need my welcome discount. Then I run sales. So they're always in there. But yeah, it was funny that 20% is the default. Everybody does 20%. The 40, yeah, 40% of our discounted orders, 20% off. That's just the big one. The median, 19.98.
Speaker B:
Yeah.
Speaker A:
That's just it, 20 is the number. It's 20%. So if you're doing less than that, that's why you're not gonna get a ton of success with it. But then—
Speaker B:
And if you're doing more than that, that was the thing that really tipped me off, is if you're doing more than 20%, it's pointless.
Speaker A:
Yes.
Speaker B:
Going over 20%.
Speaker A:
It was past 30 where like at past 30 it no longer—
Speaker B:
Past 30, it doesn't move, the line doesn't move at all.
Speaker A:
No.
Speaker B:
But even in terms of, increased conversions. Yes. So if you're giving, if you're giving stuff half off, just give it 30% off. You'll get the exact same amount of conversions.
Speaker A:
Yeah, and keep more margin.
Speaker B:
And the line from 20 to 30%, we're already getting real flat. So yeah, you might get a little bit more at 30% than you would at 20, but it's really not worth it to give up that extra 10%. Like, you should top out at 20.
Speaker A:
So the other thing that surprised me is I would think in Q4 like November, December, I should see discount usage spike. Doesn't happen. Discount usage is the same year round.
Speaker B:
Yeah.
Speaker A:
Again, it's like, okay, be smarter with these discounts. Do more segmentation, keep them limited to more promos. I think what causes this to happen is a lot of these stores have technical debt. Where like you go and look at the discounts in Shopify and you straight up cannot— you're like, you don't know what the heck you're looking at because years of running promos, you're like, well, I really don't know what these discounts are doing. Oftentimes I don't even know where they're coming from. And so how do I know which ones to turn off? Because if I turn it off, there's some like, you know, secret loyalty program in the background that no one— that everybody forgot about that was successful. And suddenly I broke that. And that example, very real. And anyone who's tried to sort through their discounts has probably had this experience where you're like, "Man, I don't even know what I can and can't shut off." And then, so then you end up in this scenario where it's like, well, it leaked to a coupon site and now everybody just gets the same discount always.
Speaker B:
Yeah, and it's insane to me to think about like, well, this is a really successful program. Look how often this discount is getting used. That's success. No, it's not. You're giving money away. Like there's a secret backdoor where you give people money. Like, that's not good.
Speaker A:
No, no, you don't want— you don't want to end up in a situation where you are paying customers to take your products. That is the fast track to failure with an e-com store. Cash conversion cycle there will eat you alive. Yeah, here's an example: a store with an Evergreen code gave back $350,000.
Speaker B:
Yeah, this code cost the store $350,000 that they could have had in revenue using another metric. Yeah. Using another way of getting people to buy. It's absolutely nuts to me.
Speaker A:
Wrong address, wrong size, add one more. The same 3 emails every single day and every one is a ticket your team handles by hand. Read it, verify it, fix it in Shopify, email the customer back. Do that a few hundred times a month and you're paying salaries to be a manual order editing tool. Our sponsor, Cleverific, has been thinking about this for years. They've run order editing on Shopify since before it was a category. Now your shoppers can fix their own orders right on the order status page. Address, size, quantity, add-ons. No ticket, no back and forth, and everything writes straight back through Shopify. And here's what's new. Order upsells that build themselves. Cleverific reads your store's order history, generates post-purchase offers based on what your customers actually buy. And it works with the flows you already run. This summer, they'll even do the whole thing for you. Mention the Unofficial Shopify Podcast and Cleverific will handle the complete setup, order editing, post-purchase upsells, all of it done for you, plus a $500 upsell credit to get you started. So stop editing orders for your customers. Let them do it themselves and let it pay for itself. Cleverific Order Editing on the Shopify App Store. All right, so one thing I saw Free shipping as a discount, one of the most successful for bumping AOV and conversion rate. So if the store does not normally offer free shipping or it's like free shipping with a high threshold, doing a sale like Free Ship Friday.
Speaker B:
That's the thing I was trying to understand. You call this out a lot. You're like free shipping as a discount, free shipping. So it's like we're not doing all orders over $100 get free shipping. —That's not what this is?
Speaker A:
—Even if it is, no, it's not. I want a discount code. I want to see it as a line item.
Speaker B:
We offer free shipping over $100. That's just the way of the world. That's just the way the store works. That is the rules of the store that you are in. Do X and we will give you a coupon code for free shipping. Oh, that's now a gift you're receiving. Yes. That's now a special little cookie they're giving you. That's not, that's just the, that's just the posted rules here. That's your special cookie that only for, we're giving only for you. People delineate that in their head.
Speaker A:
Yes. All right. And so what I'm thinking here is if you offer free shipping on your store, having it show as a rate name in the checkout is the wrong way. The right way is a free shipping automatic discount. So that way free shipping is always perceived as this is what I am saving, this is what I'm getting out of this. That's not necessarily always gonna be practical with, you know, like when you get into really complicated shipping setups, but for a store with like straightforward shipping, I would just switch to this setup. Free shipping is an automatic discount 'cause it calls it out, you know, like Shopify knows to add like the promotional displays for it. So that's a win, but then even if you say you offer free shipping at $150 and you don't, do it with an automatic discount code. Then your sale could be, hey, you know, today, Free Shipping Friday, use free ships, you know, ships free at checkout, and free shipping threshold drops to $50. And so you could do it like that, is an easy win. And again, like, I'm not giving up margin on the product.
Speaker B:
Yeah, you're giving up, you know, whatever, $5, whatever it costs you to do the shipping on the lower margin, on the smaller orders. But still, that's better than giving away 20% on the $50 order, which is $10.
Speaker A:
Well, let's say to send and fulfill my t-shirt costs me $10. That might be high. My order's $100, and previously I was giving up 20% always, but instead I say, well, it's free shipping. Okay, I gave up $10 as opposed to $20. You're gonna keep more margin doing it this way.
Speaker B:
But even better than free shipping is what?
Speaker A:
Even better than free shipping is free gift with purchase. And really it's because on the free gift with purchase, on those promos, you can set a minimum where you're like, all right, you gotta spend X to get Y. And then you could combine that with a tiered threshold. Spend $50, we're gonna give you a free, you know, drink umbrella. Spend $100, we're gonna give you a free real umbrella. And so you get way, you have consistently higher average order values with the free gift with purchase promos. And it didn't hurt conversions at all.
Speaker B:
So let's think of the scenario here. In order to get $100 in the discount method, you're— something that someone would normally pay, have to pay you $100, you're doing a 20% discount, they're only paying you $80. That's all you got now is $80. Yep. With the free gift with purchase idea, well, if you hit $100, we will give you something that is worth $5. We're not going to say it's worth $5, we're going to say it's this great amazing thing that's going to make you feel so good, but it only cost us $5. They then hit that $100 mark. All you gave up was $5. Is the cost of—
Speaker A:
you give up cost of goods sold. Yeah. Is the margin you lost on that product.
Speaker B:
So really what you got is $95, whereas with the straight discount you only got $80. Yeah. People will up their spend by way more than what the gift you're giving away is worth.
Speaker A:
Yes. To get the free gift. And the other thing, like human psychology, if I'm presenting to you, hey, today only you can get this free gift, and then I've got my orders close but not quite, even it still feels like I'm giving up something if I check out without the free gift. Like human psychology incentivizes me to want to add the item. Now with the free gift, ideally it is an item that's exclusive to free gift with purchase. Put a retail, it should have a real retail value on it, and then you let automatic discount or an app like Promo Party take care of it for you, but then it'll show a strikethrough price. And of course, the retail price has, you have to put a real retail price on it. It can't be misleading in any way. But 100%, that is, such a better option than the default. What drives me nuts about that is year-round everyone is giving away these 20% discounts nonstop. In November is when we see the free gift with purchase get used. It spikes then, like that is peak free gift with purchase. And then overall, if we did, it was of promotions that people ran, it was like 10% or less were free gift with purchase. It's just frustrating when it's like, it's staring you in the face of the solution.
Speaker B:
So you broke this down in the report across all the bat. You took every single order, 2 years worth of data, 2.2 million orders, $285 million worth of orders. All right, that's what we're working with here. And you split them up into 5 different buckets. We got the full price bucket, there was no discounts. Okay, yeah, you got the discount bucket where we're giving people a coupon code and they get 10% off, they get 20% off, whatever. You have a free shipping discount code, not the standard free shipping, you hit our free shipping threshold, now we just give it to you automatically. Free shipping as a special discount gift. And then you have a physical gift, a gift like we're talking about for you for purchase. And then you have gift plus discount, which I don't wanna talk about 'cause that's giving away too much.
Speaker A:
Yeah, gift plus discount, what? Get outta here. Not running a charity.
Speaker B:
Yeah, all right, so the full price orders, the average cart value was $112. For the discount orders, the average cart value was $116. So they did make money on it. That did increase the average order value offering the discount.. But think of how big that order would've been if you weren't giving the discount. Then the free shipping only gets $160, blows the price discount out of the water. Offering a free gift, the average cart value was $173. Yeah, look at that. Which is $60 more than the discount. I mean, that's almost 50% more. Yeah. Offering the free gift increased the order value by almost 50% over offering a discount.
Speaker A:
One incentivizes people to spend more. The other one tells them how to spend less, but they both are like, it's a discount promotion. Isn't that strange?
Speaker B:
Yeah. The discount is we're helping you save money. The free gift with purchase is, hey, buy more. Yeah. And they feel good buying more.
Speaker A:
We're incentivizing you to make a larger purchase. Yeah. Yeah. And it's, you know, the whole— it's called quite upfront about it. Yeah, spend X, get Y.
Speaker B:
So I mean, yeah, at full price you're making $112, on a free gift with purchase you're making $173. What am I missing here?
Speaker A:
Just do it. Because friction. It is so easy for me to just give away money.
Speaker B:
Yeah, I just make the coupon, you just click a button and you get a discount code. Fine. Yeah.
Speaker A:
Whereas your conversion rate went up. Okay, great.
Speaker B:
Yeah, yeah. What's— well, what's my— so what's my What's the gift gonna be? Well, I gotta figure that out. How much does that cost me? I gotta source it. I gotta worry about packing that into the orders, all that other stuff, picking and packing it and all that. Whereas the discount code, click the button, now there's a discount code. Okay, goodbye.
Speaker A:
It became the default because it was easy, not because it was right. Yeah, hmm, interesting. So I mean, my takeaway here is like, hey, first, number one, audit your discounts. There's a report in Shopify that will show you and start, you know, cleaning things up and clearing house on this. You should not be discounting 100% of the time and certainly not the same volume of discount year round.
Speaker B:
And not discounting to both new and returning customers. Yeah. They just both get the same thing all the time?
Speaker A:
Seems that way. It's like, okay, that like if this is what we're doing, just turn discounts off entirely and reprice everything 20% lower. Well, you can't do that.
Speaker B:
Because that's the JCPenney issue.
Speaker A:
Oh, not even like the retail price could just drop now because it has never actually been sold for the real price. But then beyond that, you know, free gift with purchase, man, like that is the clear winner here. But no one does it. That one, that's the one that people made seasonal.
Speaker B:
Yeah, isn't that weird that they're like, okay, that's a November thing. Okay, it's Black Friday, so we got to whip out the big guns, the things that really make money. Only this month though. The rest of the year, we're not gonna do the things that make money. What?
Speaker A:
And also it turns out a lot of the brands that are like, we never discount, they discount. They're just not upfront about it. Yeah. Yeah, that's what I discovered in this data as well. Not naming names, but no, it was, I mean, this was, it was fun to pull this data. It's fun to, you know, find this stuff out and have actual numbers. The pushback on it is, Well, I mean, it's limited to a sample of our 10 client stores that are on retainer with us and pay us for analytics and strategy work. And so like, all right, it was easy to leverage this, but what would this look like if you had even more data? If you had stuff from other countries or other languages? I don't know how it would change. And with more data, you could figure it out like per vertical. Whereas here we're lumping everything together for the sake of—
Speaker B:
I think it only being 10 stores definitely hinders us. And we have no one that's a massive commodity store. And think about subscriptions too. There's really only one store that has subscriptions.
Speaker A:
There's an excellent variety in here. Yeah. It's a good variety.
Speaker B:
I mean, it's definitely, it's not like it is now scientifically proven, this is better., but it's definitely making some side eyes emoji at being like, why don't you check that out?
Speaker A:
It is direction— I would take this as directionally accurate. Oh yeah. Yeah. So, but yeah, I think, you know, I think optimizing for AI, answer engine optimization, by the optimization for that is number 1, you are on Shopify. Yeah. Number 2, fill out the product listing.
Speaker B:
Number 3, you're done. Yeah. There's no third step.
Speaker A:
Great. You know, if you can get some Reddit comments that recommend you, that would be it, but good luck. Don't get your account banned, right? And then on the discounts, it's like, man, if you say, oh, I'm on a Shopify store, great. You probably need to go review your discounts and stop handing out 20% discounts.
Speaker B:
Yeah, that's the first thing is you gotta close all the leaks in the boat. Yep. All the discounts should be one-time use codes that expire.
Speaker A:
Yes. That was the other thing is like having a single-use code that just guaranteed it would be leaked and used forever.
Speaker B:
Yeah. And then the second is never, ever discount more than 25% ever. Yeah. Unless you're, unless you're literally liquidating and just being like, get this out of here. Yeah. With this.
Speaker A:
Yeah. Clearance. And I just need it out of the store. Yeah.
Speaker B:
But in terms of the best overall uses of your dollars is not discounting. It is either offering shipping as a free gift or offering a physical gift as a gift. Yeah. A discount is just something that's there that they already have in their hand when they're coming in. You want them to be on your store and while they are on your store, you give them a gift and that engenders the loyalty and engenders the purchase. It does create loyalty. Not, I have a special coupon and so now I'm gonna go and buy this thing that I wasn't gonna buy before. 'Cause their loyalty is to the coupon then. Their loyalty is not to the store.
Speaker A:
You're right. Yeah, absolutely. I think the human psychology of promotions, the thing people have to think about here. Yeah. But no, I mean, that's the data. The other one worth calling out was people are worried about bot traffic. It was in our data, it was about 14% were bots. And it interesting, it last November is when it peaked. And so I think, you know, that bot traffic is part of also what like may drive conversion rate down for some stores.
Speaker B:
I think that's definitely part of it.
Speaker A:
Yeah, you know, like you're— we're changing the denominator.
Speaker B:
Yeah, the bots aren't buying. And so if you increase traffic by 14% and none of those— none of that traffic buys because it can't, your conversion rate's gonna get worse. Yeah.
Speaker A:
Well, it's funny, in the data, it was a tiny percent of the bots did buy. I'm guessing those are false positives. Yeah, it's gotta be. That those were actually people, not bots. But yeah, I mean, so I've published both these reports. They are linked to from the show notes, or if you follow me on social media, I'm sure I've been posting them a lot recently. But yeah, check it out. And of course, promo party and I don't have a call to action.
Speaker B:
Well, you gave the calls to action.
Speaker A:
Yeah, go down to the show notes, click the report, check it out.
Speaker B:
Yeah, read the reports and get your product listings right.
Speaker A:
Get your product listings right, yeah, get those category metafields filled in.
Speaker B:
And close up all your old discounts and stop giving them away.
Speaker A:
Perfect, yeah, all right, we'll end it there. Thank you, Mr. Rita. Goodbye.
Speaker B:
Goodbye, like AOL. No, I got it.
Speaker A:
Hey, before you go, I was hoping you would check out our new app, Promo Party Pro. It is what I want to be the single best, easiest way to run a free gift with purchase promo on Shopify. We just put it live in the App Store. We've got less than 50 users. We want your feedback. So if you need to run a free gift with purchase promo in the near future, install it, try it. There's a live chat. I check that all the time. And so if you have any issues at all, or any suggestions on how we can make it even easier to use, let us know. We're happy to help. If you wanna try it, search PromoParty in the App Store. PromoParty Pro's the app. Give it a shot. It's got a free trial. Thanks.