S14 E5: How AI Is Rewriting the Rules of Ecom Marketing (with Liam Millward, Co-Founder and CEO of Instant)

0:00Nik:Welcome back to Limited Supply, the podcast where we get deep into the tactical and strategic side of e-commerce, digital marketing, and building consumer brands. I'm your host, Nick Sharma. I've spent the last nine years building, scaling, and investing in brands. And through this show and my weekly newsletter at nick.co slash email, I'm here to share everything I've learned. The wins, the losses, the experiments, the tactics, and the insights, all so you can unlock your next$100 ,000 in revenue. Today's episode is a good one, but before we dive in, let me tell you about our chosen sponsor for this week's episode.

0:34Nik:Think about your best friend and how they shop. Totally different than how you shop online, right? You should never be getting the same cart reminder email. That's where Instant AI comes in. In the last few weeks, Instant rolled out AI flows for over 500 brands. The momentum is crazy, and one of my favorite brands tripled their email revenue in just 60 days. That's one of hundreds of wins they've had so far. Here's why it works. Instant can immediately understand who's on your site and what they've been looking at. The discount, the copy, the imagery, the on-site promotion, all of it is pulled dynamically into personal email flows.

1:07Nik:Then it sends it at the exact moment that that shopper is ready to buy. BFCM is behind you, and now's the time to build for your biggest year yet. Old and lame retention emails are so 2025. Book a demo by December 31st and get 50 % off your first 60 days at instant.one.sharma and set up your brand to dominate 2026. All right, Liam, so you're back on the podcast. Today, we're going to talk all things AI. And, you know, AI has evolved so much. Everything from, you know, well, I think the thing is like AI is not a product. It's just a it's a new tool that I think everybody has the opportunity to use.

1:49Nik:And I feel like those who do not figure out how to use it will find themselves being consumed from it. and so you know yesterday we talked about AI at a conference as well but you know I guess before we go into all this why don't you give a quick background of yourself at the intersection of AI why is it fun to have you on and talk all things AI yeah well thanks for having me back I mean we've had an incredible podcast previously and we've spoken on so many stages and and together in your newsletter, et cetera, at this point. So thanks for having me back on. I mean, AI is, if you're not adopting it in your business now, you're already behind.

2:27And one of my topics that I speak about on stage a lot is that if AI isn't driving at least 25 % of your revenue today, whether it's across email, SMS, retention marketing, you're already failing. And I think that AI is still being spoken about, particularly in the world of e-commerce and brand founders, as a hypothetical. It's coming in the future. It's not here yet. It's really cool. Or people think they're using AI because they're addicted to ChatGPT. But then there's this completely other side where AI is real. And if it's not such a big part of how you're thinking in the operations of your business today, you're already behind.

3:08So excited to cover some of that. In terms of Instant, I mean, started the company when I was 17. Um, we work with thousands of brands today, started the company when, uh, I was back in Australia. Um, now spending majority of my time in, in New York. Um, and I think that we actually started as a shop pay competitor. Um, I didn't know that

3:29Nik:until recently. Yeah. To Shopify. That was our, that was our biggest product. We, um, helped brands with one click checkout, um, remembering their shoppers every time they came back to their website. And then eventually we decided that Shopify was the beast that we didn't really want to compete with. And so we used that technology of being able to remember a shopper every time they came back to your website and bring it into the marketing world and help a marketer unlock their anonymous shoppers that are visiting their website. 98 % of people that visit your website, you have absolutely no idea who they are.

4:04And so we uncovered a huge portion of that, enabling brands to trigger email marketing. And alongside that, I've been on the podcast before talking about audiences and instant audiences and helping brands unlock that channel. And that grew so quickly. That product, we went from zero to 500 brands in 10 months, which was wild. But out of that and where we started to invest in AI was we were building all of these optimizers because across all of our portfolio of brands, email marketing was something that just had zero investment, you know, hadn't been updated in a year, was seeing like declining results.

4:43And we built this tool that helped the brand optimize their flows and campaigns. And eventually we stood back and went, hang on, why don't we just do it all for you? Um, you know, we've already identified the problems with these emails. Why don't we just fix it with AI? Um, and that's our biggest product today, instant AI, which helps brands to completely, um, uh, automate their email marketing from end to end.

5:07Nik:And that's actually doing all the generative copy, pulling imagery in. Are you making these emails from scratch? Are they just templates? We make them from scratch for every brand. So we had a huge fashion brand in Australia, Fate the Label, Simile, Karen Kane, here in the US, Third Love. And they're all up and running in like 24 hours. They have a conversation with us and they're live 24 hours later. and the way we do it is you input the URL of your website, we connect to Shopify and we basically analyze your website. So we look for your tone of voice based on the copy on your website, we pull all your brand colors, images, logos and then we use that information to create one-to-one emails to every shopper.

5:57So every shopper receives a different email, no two shoppers receive the same email twice um and for those brands what they saw in the last three months uh with their ordinary outdated flows uh we generated for them in the last 30 days um with instant yeah so it's pretty crazy i feel like there's uh there's so many flaws in the way

6:16Nik:that modern like the way email is handled today right most brands will maybe have one person internally who's managing email but like they're pretty much if you're lucky yeah and they're pretty much drinking from the fire hose, trying to get the next campaign out, you know, managing a copywriter, a designer, putting it into Klaviyo, you know, maybe exporting audiences for the paid team, running reports because the Klaviyo reporting is not that great. Or you have an agency and the agency's goal is to just churn out, you know, a high volume of creative, but then they're not usually going back and looking at flows or updating campaigns or testing new dynamic blocks or figuring out, you know, what A-B testing, you know, layout works better than the other.

6:56Nik:And, um, I mean, to your point too, like being able to pull from the site, you get a lot of other signals that you just can't do at scale. Uh, cause you know, those are one to many versus this is sounds like it's all one to one. Exactly. And we, um, we, a common misconception amongst agencies, um, or brands when it comes to AI is stealing people's jobs. Um, and it's just not true because it's never been possible before. And it still isn't possible today for either an agency or a marketing manager to sit there and send a unique email to every shopper that abandons your site. And so it's not a war against this mindset of having no agencies in your business.

7:42There's still absolutely a place for them. However, emails that or attention marketing that is unique to every shopper that visits your website is it converts wildly better. And an example of this is today, every single retention flow that you get, no matter what website you get will be, hey, Nick, saw you abandon your cart, come back to purchase. There's like a standard playbook of 16 minutes. It's a standard and everyone just goes, delete, delete, delete, never has any cut through.

8:12Nik:No discount, no discount, 10%, 50%, 20%. Right, because that's all that's available. Right. Whereas now and like one of our, you know, really large brands, July Luggage, you go on their website and rather than it saying basically any brand that uses Instant AI, there's thousands of them now. It starts to pull in copy into those emails, like the name of the shopper, the item that they viewed, maybe for the luggage, you know, it's telling, it's reminding the shopper. it might say like, come back and check out the pistachio dark green travel case. This is your perfect travel case for your next trip.

8:50Um, 30, you know, blah, blah, blah, leaders, um, uh, free shipping, like the emails auto update based on the updates that you make to your website as well. So sometimes you're doing free shipping. Um, and so that copy is automatically, um, injected into the email.

9:08Nik:It's like an agent is constantly going, reviewing the site almost like i was always up to date yeah and like somebody's sitting on the site watching this viewer come in see what they're doing and then tailors the content basically just to them yes exactly it's um it's very hard to explain like on a podcast yeah such a visual product totally but when you see it i mean for me the whole instant thing was visual like when i saw the demo yeah that's when i was like whoa this is crazy and then the revenue results just back it up as well right um and that's what i think excites brands and us most yeah i mean we see usually like a 2x base or double lift um but what do you think are like you know there's been all these shopify sass companies that have been doing ai and personalization for years what do you think is actually now possible with ai whether it's generative or you know in the personalization world that wasn't possible a couple years ago i think copy personalization really is just the surface.

10:01Um, and, uh, there's a lot of companies, a lot of, um, opportunity with copy generation to really increase conversion. If you're not using AI to generate copy, unique copy to every single shopper, no two shoppers should receive the same email twice. Um, you're not doing that today. You, you are, um, already, yeah, I mean, you're, you are so far behind. It's crazy. Um, and the, the brands that invest in that early are the ones that are clearly winning. But there's also so much other opportunity that really moves the needle. Like one of our biggest features today is smart coupons. Ordinarily, you know, it's the standard to do kind of what you mentioned before, right?

10:48Every shopper gets 10 % off, no matter their behavior, no matter their shopping history, no matter the time of day. Whereas now a brand can set a minimum order value and a maximum percentage and let the AI do everything in between. Some shoppers will get 12 % off. Some shoppers will get 10 % off. Some shoppers will get 0 % off. Some will get 20 % off. Every shopper gets a unique discount based on the minimum percentage that the AI thinks to convert that shopper. And so you're converting more shoppers. You're also predicting margin and profitability as well in a completely automated way. Um, another really cool thing with AI, um, is send time optimization, um, in your emails today.

11:30Like if you look at any report across the internet, um, any case study, any downloadable, any email that these SaaS companies put in your email inbox, the number one thing that they always say is to send emails two hours after a shopper abandons their cart. Um, and the reason why they say that is because there's just no, uh, there's just no alternative. I mean, what's the alternative, right? And however, today, you should be sending emails based on when that particular shopper is most likely to convert. Not a blanket rule, two hours to everyone. Like Nick, he loves shopping at 10pm at night when he's lying in bed.

12:07So let's send him an email then to remind him to purchase. You know, this really busy mom, she really loves to purchase at 8pm at night when the kids have gone to bed. and so based on her previous shopping behavior let's send her an email then versus just two hours

12:23Nik:it's almost like having like a little team of agents and they basically sit on your site and they see somebody come in they get the email or they figure out who the person is they try to connect them with an email then they watch what they do on the site and then they figure out what they know about this person from before and then they craft a personal email and they fire it off except that happens in like what 30 seconds easily Um, and even down to the subject line, um, uh, it's again, like you, you can have all these little features, um, but it's never been worth a marketing manager's time to sit there and go, shit, what is the subject line that is best going to convert?

13:01And what happens is a marketing manager will sit there once and they'll craft the subject line. Um, and you know, that's the subject line that they stick to forever. Totally. Um, whereas these AI agents now just consistently experiment with the best subject lines, the subject lines consistently change based on the time of day of that shopper consistently change based on, uh, you know, the sale that's running on your website. If you're running a birthday sale, that subject line now includes it. You get rid of that birthday sale and now have free shipping.

13:31Nik:It includes a lot of like micro optimizations that you can now make because the technology enables it. Exactly. Yeah. And the biggest feedback we get from brands is, all right, email marketing is 20 to 30 % of our business. Basically, nobody spends any time on it. And I think the biggest reason for that is because, honestly, if you look back, what innovation has there been in email for the last decade? It's been the same advice, the same percentage of your revenue for decades. And so I don't think it's a matter of brands not wanting to invest in the innovation of email. There's just been no innovation.

14:09Yeah. Whereas that has changed.

14:11Nik:There's been, yeah, there's been some innovation I feel on, on the SMS side, but there's been very little on the, on the email side. Yeah. And also there's been no, um, well, I mean also, yeah, it's just, uh, I don't know why there's been no, no innovation to be honest. I think, I think there's like, there's a few players in the space that are just happy with being what, wherever they're at right now, maybe. absolutely i mean um i think that brands that if a brand could increase their their revenue from from email they would be ecstatic because it hasn't moved in forever yeah don't cost them anything you're already getting these shoppers to your website so you're already spending a shitload of money in meta ads to get these people to your website in the first place they're landing there you may as well be targeting them uh to bring them back yeah um if you can increase the amount of shoppers that are identified on your website and increase uh or improve the emails that are going out to um to bring these shoppers back to convert you win totally um and like i think the revenue results just speak for i mean one of the brands that we work with is higher dose right like 3x in 60 days yeah crazy more than 3x but yeah nuts and from a relatively simple or relatively low time investment.

15:29Nik:Where do you think AI is finding itself in other places in the marketing stack? I think that... Do you think the same way that you're able to generate emails on the fly based on the signals or data points you grab immediately? When is that coming to a landing page where as soon as Instant identifies you're the person coming to the landing page, it immediately builds the page. I think that stuff is coming. I really do. I think that today we definitely focus on kind of the retention piece. But where I really see AI going most when it comes to marketing is what I was just speaking about around, um, these really, uh, consistent optimization, consistent experiment.

16:23Um, and one of our biggest features is, uh, called instant labs. Um, and it's just, you know, unless you're a brand that is willing to sit there and do these, you know, really frustrating AB tests, testing one, you know, two things at a time, um, nobody's interested. Um, and then you do the experiment. And it's like, well, of course it generated more revenue because of the time of year. And so it's just so hard to get an accurate answer. And so you just stop doing it. Um, and the brands that need to do this most are either really big brands that have, you know, so much internal complexity or really small brands who just don't have the manpower.

17:03Um, and so I think AI now gives a brand the ability to do these micro optimizations consistently without any human involvement um consistently figuring out the best subject line formula we spoke about this on stage yesterday where a women's fashion brand we see having a an emoji the shopper's first name comma and the name of the product being the best winning formula um in terms of a subject line but that subject line formula significantly changes during black friday cyber monday right um which significantly changes to a store that heavily focuses on men's products or, you know, it has a different persona.

17:46And so these micro optimizations on a store by storefront is very interesting rather than these blanket, I guess, blanket rules or advice across the entire industry.

18:00Nik:Yeah. Brands like Third Love, Liquid IV and NeuroGum all swear by instant. They've tripled their email revenue and some like Karen Cain even see a 5x increase. Why? Because AI can build hundreds of versions you'll never have time to make. While you're running your brand, Instant is running the experiments, figuring out what works, and then continuously optimizing to keep conversions growing. BFCM's over and your next growth run starts now. Grab a demo by December 31st and get 50 % off your first 60 days at instant.one slash Sharma. if somebody wanted to like you know we've basically evolved from the generative ai to the agentic ai somebody wanted to start using these you know building agents and you know everybody talks about getting agents to replace routine tasks even if it's something as simple as you know generating a report about email performance for the last 30 days or analyzing that and giving opportunities to do better.

19:01Nik:How can people start with building things like this? It's interesting. I think that there's so many tools out there now, like Lovable, Replit, etc. You can really start to experiment with different tools. I guess we're still in the phase where people are learning. I think that if you're not learning or becoming very sharp on AI, it's basically you're becoming illiterate. And if you're not adopting it into your business or as a leader, you're not pushing people in the business to adopt AI, you're already behind. I think that's the biggest thing that I, you know, talking to merchants today, it's always the senior leader in the company that is pushing back because maybe they're scared of it.

19:49They don't understand it. They don't experiment with it. And so it's all of a sudden the leaders of these businesses actually holding the business back. And so I think, you know, using those types of tools, you know, there's also so many other tools out there like analytics platforms are really starting to bring AI into the business. Shopify is heavily leaning into AI if you're not using those tools. Yeah, with Sidekick and whatnot. Which is huge, right?

20:19Nik:I feel like, you know, a few years ago, I was getting called into like the boardrooms of like Tom Ford, Diageo, Bacardi, and these guys were like, we don't know anything about direct to consumer. Yeah. We need you to come teach us, you know, and maybe like we hire you to basically educate our whole staff on how this works and, you know, do an audit of what we're doing wrong and what we're about, what we need to change. I feel like that same type of wave is about to hit on the AI side as a lot of these like senior leaders are like, yeah, we know we don't use AI and they're reporting that up and the board's like, what the fuck do you mean?

20:57Nik:You don't, you say hi. Yeah, that's what's happening. Yeah. Have you gotten calls like that? I feel like you're like the person to call for that. A huge, we always lean into, I mean, we have so many features across AI now that it's hard to jump on a podcast or a particular piece of content and explain it all. Yeah. And so I think that we get so many, yeah, I mean, we had 300 brands in the last six weeks alone um moved to instant ai which is crazy um and a big push for that is not the hardcore sales you know you need this you need this like we don't cold call anyone um it's brands coming to us going well we just received this um this wildly unique email from another brand that i was shopping on we need that on our site um or i was at this dinner with x brand last night and they said that they had three times their revenue on email, unheard of, haven't focused on email in years, we need instant.

22:00And so I think there's word of mouth amongst AI tools. You should only follow, you know, adopting an AI tool if you hear about it organically because they're the best ones.

22:10Nik:Yeah, there's a lot of apps, a lot of advertising garbage in the AI world. Absolutely. And look at the quality of brands using them as well. I mean, there's plenty of brands that, And I think this is going to be the real shift over the next six months is seeing where individuals, businesses, D2C brands, what tools become very sticky. I mean, you look on Twitter X and majority of these AI tools are like, yeah, we've gone from zero to$100 million of revenue in six, 12 months. um i think that that that a lot of that will be very um you know interesting to see if those businesses actually i feel like if they scaled at if they scaled that fast it's going to be easy for more people to just replicate that do it themselves whereas the the really good tools um i i think will start to pop up over the next six to 12 months where these these really large businesses are relying on it um to to run their majority of their revenue um and today a brand that uses instant ai we are accountable for about 20 to 30 percent of their entire revenue their site revenue wow um driving that instant so you know it's not these i would be looking for you know there's plenty of tools out there that you know can drive like two percent of your revenue three percent of your revenue is a nice to have charging you thousands of dollars a month yeah but you need to look for those tools that are like you know straight in um and really moving the needle on your business totally where do you think um like cmos today can spend time incorporating ai i think that i would be focused on incorporating ai so i can focus on the most important things in the business.

23:56Um, for example, you know, there's been lots of innovation in SMS marketing. Um, like AI should be running your SMS marketing today. Um, there's a lot of innovation, uh, in, uh, in, uh, retention experimentation. Um, AI should be your experimentation agent today. There should not be a single person in your business that is, um, doing one-to-one AB tests, um, and experimentation. AI should be doing that. There is, you know, in terms of email marketing, AI should be running your email marketing from a retention perspective almost entirely. A human should be going in and maybe having a small part of creative control to make sure it's on brand.

24:42Maybe adding some things to a campaign.

24:45Nik:Have you played much with Nano Banana? I haven't. No? No. What is it? It's one of Google's image video models. Okay. It does a phenomenal job of removing things, swapping products out. I mean, there's a couple of GLP One brands that I'm aware of that their best ads across the ad account where they're spending$100 ,000 to$200 ,000 a day are videos of women holding products or taking something out of a box. It's all made by Google using renders and like stock video and then putting them together. So good. And I think, I mean, the image generation was something that was behind for a long time. For a long time.