Moats for GOATS: Building sustainable and defensible Shopify offerings in the age of AI

Gavin: Thanks, Trudy. Thanks, Taylor. Really happy to be here. Really excited about it.

For those who don't know me, my name is Gavin. I've been around in Shopify for a while. I've been agency side, I've been freelancer and I now run apps, although we call them a platform because that sounds fancier. Yeah, always happy to chat Shopify, so definitely grab me after the talk. So the topic for my talk today or the title is Moats for Goats, Building Sustainable and Defensible Shopify Offerings in the Age of AI. But really the subtitle and the subjects for that is, "Are we fucked yet?" And I think you know there's a lot of concern around that in the ecosystem, certainly the conversations that I've had in the last couple of days at .dev, especially in the app ecosystem where at least the narrative is that people can vibe code an app in a couple of days, get it into the app store, you've got copycats coming in left, right and centre and at the same time you also have Shopify pushing things like Sidekick being able to build a custom app in a couple of prompts and I think there were probably a few clenched cheeks when Atlee was sort of demoing that in the keynote on Tuesday morning.

There's obviously a big concern about are they going to get replaced. I'm an optimistic person though, so I think Betteridge's Law of Headlines says that if you ask that question, you have to answer it no, and I do agree with that. I don't think that we are quite there yet.

The thing that I thought is if what you ship is code, if what you think you sell is code, then you're in trouble, but that's always been the case. From a business perspective, I don't think there's really any difference between someone vibe coding an app and shipping it versus what they could do five years ago where they could hire an offshore team to try and clone your app and get it out the door. The time scales are compressed, obviously, but the fundamental difference and the fundamental difference to what your business strategy is is kind of the same.

So again, it's not the code that you're shipping that you're selling, it's something else. And all I want to talk to you guys about today is what is that something else? And I've just got six things that I kind of think are different modes that people have built, are building, and that you can potentially build.

I'm sure lots of people in this room are thinking about this a lot. But the hope is just by talking through a couple of examples, maybe that triggers something in your mind and gets you thinking about your apps a different way. And if at the end of the talk you can't think of anything that your app fits in or a category that your app fits into, then maybe start thinking about that.

The first one is brand and trust and I think this is very obvious, but it's also the thing that is always going to survive AI, I think, no matter how, where it goes in the future. And it's also the thing that's the hardest to fake in a world of AI, so I think the value of this becomes higher and higher and higher. I think Sandesh spoke about it really well when he was talking about their journey with STOQ, but another couple of folks that I think do this really well in the ecosystem, I mean, I had to say, Klaviyo does it well, but they've definitely had the time, the money to build up that reputation in the ecosystem and they're very much in the position where it's no one ever gets fired for buying Klaviyo type of vibes.

There's a hand at the back because you can't hear me, or no, you're just stretching. That's good. Stalingbot.

Another app that does this really well, I think, is Mechanic. And if you guys know them, they've been around for ages, but I think they're built trust in a different way by being friendly and nice and super helpful and considerate of developers versus being a giant billion dollar BMF. But they've just invested a lot of time in thinking about how they can make the experience really great for developers and merchants and things like that.

And I'd say of all the things I'm touching on here, this is the one that everyone can and should be, and need to be thinking about. Another mode you can build is around data. So what information do you have access to that no one else does, that an AI is not going to be able to replicate?

Two examples of this that I think are kind of interesting, one's Fairing which is a post-purchase attribution tool. They have not only the data that they've collected over the years, but they also have the data data of everyone else on using their platform and they can use that to surface better results. So if your app has anything to do with data, if it collects any data, start thinking about how you can start leveraging that and making that something that isn't as replicable.

Gorgias has this as well. Obviously they have their ticket database, which again is not only useful for each individual specific merchant, but they can also use that data and train their AI agents across the the whole dataset which makes things a lot more valuable. So you also get the network effect for them as well.

This one's quite relevant I think for us 'cause I feel like this is the mode that we have. So for anyone that doesn't know, we run an app called Submarine. It's kind of an enterprise focused subscription service plus some other stuff.

But what we find is we have APIs, we have code, We have a UI, we do stuff, we deliver value. But really, a lot of the time when people come to us, merchants, agencies and things like that, what they're buying is not that code, it is the expertise that we have. We've been around for a long time, we know how to make really complex stuff work, and a lot of the time the reason that we're getting pulled into it is because people like, "How do we get this to work on Shopify?"

And we can help them out with that. So if there's an element of a service element to your app or you have some specific domain knowledge that no one else does and that's why you built the app in the first place, I think really lean into that and lean into the personal expertise part of it because again that's not something that can be replicable that easily. The other example I have here, this is one for the agency people, which is an agency He's a friend of mine in Australia, Bedford Lane.

And Alex, the founder of that, is actually... I did check, he's all right with me saying this. He's getting rid of his agency, basically, and he's throwing away time and materials, billing, he's throwing away paying for outcome.

And what he is now saying that they are selling is their expertise in running their internal AI system. So he's basically renting out their AI models to merchants, and that's how they're phrasing it and pitching it. And again, that's, you know, from a merchant's perspective, they feel like they're getting the benefit of AI, they're getting the benefit of that expertise, that's what they're buying.

And at the end of the day, he's expecting that they're gonna be charging and billing the same, if not more for similar, well, better outcomes for everyone. Another mode, slip. I use this specific word because I think there's a Paul Graham essay where he talks about the folks who started Stripe.

And when Stripe got started, payments in general on the internet sucked for the developers. It was, I think it was PayPal, and that was kind of the only option you had. And everyone knew it sucked, and everyone knew there was a crapload of money to be made if you built it right, but no one did it, because it's an incredibly hard schlep.

There's just so much compliance and regulation in different countries, and to make it work. So if you are willing to put the time in and do that sort of stuff, again, it's not the sort of schlep that an AI is able to replicate. These are all, I don't want to say Swap is a boring company, but TaxJar is a boring company because tax is boring, but again, who's attacking that vertical?

Who wants to deal with the millions of layers of regulation that you need to do when you have the US with county level tax and things like that? we've got incredible regulatory requirements and things like that. So if starting an Affirm clone isn't like something you're going to vibe code over a weekend and then Swap like cross-border commerce and all of the duties and the regulations and the license that you have to have. So it's a hard slog, but if you're willing to make it, then that is one of the things that you can use to make your business defensible. Risk shifting. And I And I think this is really powerful if you're thinking about the risk of a merchant wanting to replicate your app.

If the whole reason they're buying your app in the first place is to get away from doing something themselves, then they're going to keep paying you for that. Like Rewind's the most obvious example there. Like who's going to buy their own Rewind app?

It just doesn't make sense. So if your business model is, hey, we're going to take risk away from you, we're going to make your life easier, then people are going to keep paying you for that no matter what the code looks like. And then the last one I had, and this is one that I feel is counterintuitive and one that's counted or tutored a lot of advice that I have ever given, which is people should charge more.

But I think it's a really interesting example of a moat. And that's Judge.me. So, Judge.me being really interesting, I think, because their cost is essentially zero.

So, I mean, their business model is $15 a month. Flat, you get access to everything. But for a business, that's $0, like a round down to zero.

You'd spend more on the tokens trying to replicate that. And so, why would anyone do it? So, I think it's a really interesting angle.

I don't know if that makes people a bit nervous because it implies that there's downward price pressure, but I think it is genuinely a really interesting, a really interesting angle to take in the age of AI. So that's kind of like the six that I came up with. I'd be interested to hear if anyone else has any other ideas on it.

And again, hopefully as we've gone through this, like I put up some specific examples, but hopefully you're thinking about where does my app fit in there? Do I have one of those things? Do I have something else?

And again, if not, maybe start thinking about it because I think you're gonna need that over the next couple of years. Thank you.

MC: Thank you. Okay, we're ready for questions. Not even from Jordan.

No, not even. (laughing) Oh, I've got one in the back. Brian.

Do I get a hat? Catch, buddy. Brian.

Audience: So I didn't really hear you say anything about features. And Sandesh, if you had like a conversation about, you know, your core feature being compounding effort. And so, can you speak a little bit about features?

I'm going through something right now with our own app where we started off with a really simple idea. And I did get the blueprint from Matt DeSousa. So I followed that in the beginning, but then I got excited because the app was live and we're now ready.

And then the features kept coming because there was certain down times that we had. And my engineer, Nick, he was like, Oh man, you know what we should do next. And he just kept building.

And next thing you know, I do feel like there's a bit of a moat around the idea of this really simple concept. It was like internal documentation inside a Shopify. So like great for developer handoff or whatever.

But now it's become so much more of like a productivity tool. It's almost like a sauna inside a Shopify. And I was like, I had no idea it was gonna be like that.

Maybe that created a little bit of a moat because of the fact that like, if you're gonna use that many tokens or something to build all these different features, like you probably could. But yeah, can you just speak a little bit about features? 'Cause I'm probably in the wrong right now.

I definitely know that that maybe wasn't the best idea, but go for it.

Gavin: Yeah, so I would sort of say in that context, you're not, like the features aren't necessarily, the mode is your experience going through that process and having like understanding what are people, and hopefully people are asking for this stuff you're building it, but I don't think merchants aren't buying things for the features. They're buying it to solve a problem and they're buying it because you have the expertise in having built the best thing that can do that, I would say. At the end of the day, you have a list of features on your app page, but that is probably something that someone else can come along and go, "Right, I'm just going to smash this out."

The value of features in and of themselves is the same as the value of code, which is trending very quickly to zero. And I say that as someone who writes code and loves code. But yeah, I think that's a good example of what they are buying is your expertise, is what the app is actually doing for them, and the way that you're differentiating it is you've been doing it longer, you've been thinking about it more, and you're not...

So I think that's how I think about it.

Audience: do we have some advice for new apps or young apps that enter this territory? How can we build the brand and trust from the beginning?

Gavin: Yeah, if it was easy, it wouldn't be a moat. But I actually think what Katie was talking about before about building in public and being a bit more open, and I think she made a really great point about how... There are lots of conversations in the Shopify ecosystem, And I think people are generally, these sort of events, I think people are very happy to share knowledge.

And that's one of the things that I think has always been great about the ecosystem. But yeah, do we have a tradition where everyone's publishing a year in review? And that's kind of a thing that the ecosystem does know.

But that's the sort of stuff I think could be really interesting for someone who's just starting out because you obviously, you don't have 20, like it'll take 20 years to get 20 years of brand trust.

Audience: I've got a question, Gavin Gray talk. I'm kind of, when I listen to what you're saying, it made me think of this James Dyson quote, like that you need a passionate anger for something in order to fix it. And I guess it's kind of a little bit of an insight, but also a question to you if you've experienced this, because as someone who talks to clients regularly and hears their gripes and pains, it's actually part of my discovery process when creating something on the agency side.

I feel like I hear a lot of these gaps and these problems that they're trying to solve, but I don't create apps. So I have though been in these conversations where like Trudy will call me and say, "Hey, what do you think of this new feature? Like do you think that'd be helpful for your clients?"

Or even in Vinoda did the same thing and I've talked to other people. Do you get into these scenarios like as an agency and an app solution guy where you're feeding your own pipeline of ideation? And if you're not doing that already for those who are out there, definitely make an agency friend who talks to 20 clients every week and check in with them if whether your thing is a great idea that directly influences what that person hears, of course, or check with a couple of people just to make sure.

So back to you on the question and addressing that.

Gavin: Yeah, I think there was a question in there, was it?

Audience: There was. Agency. How do you feed the ideation process? Yeah. do you see value in that relationship would be the actual question to you?

Gavin: I mean, a million percent, I think this actually ties a little bit to the expertise or maybe even the schlep. Like, what we built came out of what we did as an agency. And I was freelance for a couple of years.

We run the agency for ten, well, maybe eight years before we pivoted to products. And there's no way we would have understood the enterprise subscription market and large agencies and how large merchants think and all that sort of thing if we hadn't done that and done that for a long time. And as a result, I think, and that's again, that's experience that isn't replicable.

So someone can look at what we've built and try and build it, but they don't know why we did it that way or they don't know what our relationships with these merchants and these brands are. So yeah, I think some people have probably done this successfully, but in general I would would say if you come into the Shopify ecosystem and you go, I'm going to build an app and it's going to be this and it's going to be great and you haven't lived the pain of merchants for a little while, it's unlikely to be successful. So yeah, I think it's a really good point as well.

Like if you want to cheat code that and accelerate it, make some agency friends and see what they hate, but also you're not going to feel that pain quite the same. It's going to be reflected pain, I guess. Great question.

Thanks. Thank you very much. [APPLAUSE]