When AI can build and copy software in days, what is your startup moat? Peter and Dave argue the protection that lasts comes from customer value, not from shipping features first.
Getting a product to market used to protect a well-thought-out idea. Now development is basically a utility. An incumbent with funds and technical skill can point an agent at your product and copy it the week you launch. Peter Maddison and Dave Sharrock talk through what still holds up. Traditional moat tools like Porter’s Five Forces and network effects were built for global players like Uber, and they do little for a small startup. The protection that still works is knowing your customers’ pain points well enough to fit a narrow niche better than anyone else, then running the experiments that prove it. They also find a moat nobody expected: the internal processes that made your company successful can now become custom tools that only you have.
This week’s takeaways:
- A copied product does not win customers by itself, because people stay with tools that work well enough unless they have a real reason to switch.
- Product market fit has gone from a nice idea teams nod along to into a must-have, since understanding the customer’s problem is one of the few advantages that is hard to copy.
- Small and medium companies can turn their own back-office processes into automated tools built for a market of one, and competitors cannot buy that off the shelf.
Listen to the full episode at definitelymaybeagile.com
Subscribe so you never miss an episode.
Have a question or topic you’d like us to cover? Reach out at feedback@definitelymaybeagile.com
New episodes released every Thursday to challenge your thinking and inspire action.
Listen and subscribe:
Welcome And The Moat Problem (0:04)
Peter Maddison 0:04 Welcome to Definitely Maybe Agile, the podcast where Peter Maddison and Dave Sharrock discuss the complexities of adopting new ways of working at scale. Hello and welcome. I miss being able to do that at the start of these calls. It's always fun. Anyway, how are you, Dave?
Dave Sharrock 0:21 I'm doing well. As you can probably tell from the voice, I'm recovering from some fever or other that just came through, and a cold. But other than that, doing very well.
Peter Maddison 0:31 Yeah, a little nasally.
Dave Sharrock 0:33 Yeah, well. Adds a little intonation to the ups and downs of the conversation. Hopefully something a little bit more.
Peter Maddison 0:40 Yeah, we'll keep it interesting. All right. So what is the topic for today?
Development Is No Longer The Moat (0:44)
Dave Sharrock 0:44 Well, you and I have been talking backwards and forwards over the last few weeks around the difficulty of finding a moat to protect the startup idea. We've been working in startups for a while, either in them, alongside them, helping people do them, whatever it might be. And I think it used to be that a good idea, well thought through, brought with it, and I'm simplifying, some moat-like protection.
Peter Maddison 1:15 Yeah.
Dave Sharrock 1:16 And that really is disappearing in a world where development is basically a utility.
Peter Maddison 1:21 Yeah. And so the ability to bring the product to market isn't the differentiator anymore.
Dave Sharrock 1:30 Yeah.
Peter Maddison 1:30 Because we can bring a product to market very, very quickly. I was chatting to somebody about this earlier today. They did some work for a Martech company, a marketing technology company, and they brought the full platform to market in two months. Now they're using that for three of their clients, and it's working great, generating revenue and doing all the things they needed to do. That's just normal. Business as normal. That's easy to do. And then one of the things he was saying is that they realized it'd be really great if they had an events platform. So they went and built an events platform over a weekend. They ran a hackathon and created an events platform so they could say, "Okay, I need to run an event tomorrow, go book me the location, get all of the stuff set up, do all the pieces I need to make sure the event can run," etc. It'll go do the bookings.
Internal Automation Becomes An Edge (2:22)
Dave Sharrock 2:22 I think you're identifying this really interesting space. A lot of companies have determined, over many years and many clients and many experiences, back office processes that work really, really well. We've been involved in lots of process re-engineering exercises where, as they bring in new systems, they have to shoehorn in these ways of working. Maybe optimized, maybe traditional and needing a bit more optimization, but their own customized ways of working that work really well for the niche they're in. And now those can be automated very quickly, very easily, in a system that has no market. There's a market of one, the Martech company you're describing, because of their unique ways of working. However, they're in control of it, they can drive it. If they need an events platform, they can build that out nearly as quickly as it takes to have a conversation about what it could look like.
Peter Maddison 3:19 Yeah, when we're thinking about the moats for the stuff that they're doing, it's this: if I need a solution, building my own solution becomes a much more palatable option. As I've heard it described, it's AI sourcing versus outsourcing. You've still got to be able to manage and support and maintain and upgrade and all of the other fun stuff that goes with that. But it becomes much more palatable than...
Dave Sharrock 3:52 And very possibly a competitive advantage. I'm thinking of a lot of the small and medium organizations that we step into. What we often see is that they're hamstrung with paper copies in certain places, and handoffs, and all of the things you find when you look at process flow and where the bottlenecks and handoffs are that slow us down. All of a sudden those become something you can really accelerate and smooth over internally, right?
When Buying Software Still Makes Sense (4:20)
Peter Maddison 4:20 Yeah, exactly. Well, so if everybody can do this, nobody ever needs to buy any software. We're all going to build our own, of course.
Dave Sharrock 4:26 There are reasons to buy software, right? To get into larger organizations, all of a sudden there are the compliance aspects, and, I'm smiling as I say this, the reassurance that the services you're purchasing are going to be available 365 days.
Peter Maddison 4:40 Like Microsoft 365, for example, which is clearly a very stable platform.
Dave Sharrock 4:46 Well, that side of the moats, I think, is not really what we're talking about. There's opportunity in building internal systems, which becomes more and more of an advantage in a competitive environment. I'd say the organizations that are persisting with the old ways of working, without looking to automate, to accelerate, to really understand and make that their own, there's a problem there.
AI Copycats And Competitive Response (5:15)
Peter Maddison 5:15 Yeah, but that's what we've been talking about for decades. I think where the interesting piece comes is that yes, you could go and copy something. You can point your agent at it and say, "I want something that looks like that. Go build me one of those." And it can go create and build out that capability. I think there is a point at scale where you hit, "Well, now I'm going to have to maintain and operate this thing." Of course my agents can help with that too, but if you don't have somebody who knows enough about it to do that, that could be a bit of a...
Dave Sharrock 5:55 But I think even that is becoming less and less of an issue, because eventually that's going to be something that's black boxed and...
Peter Maddison 6:01 Yeah.
Dave Sharrock 6:01 At least, you know, we're not talking about...
Peter Maddison 6:05 The environments that we exist in. And as Microsoft proved this week, it's very easy to totally screw things up, no matter how good you think you are or how many agents you have. Things go wrong.
Dave Sharrock 6:20 Yeah.
Peter Maddison 6:21 And certain problems, and we'll see the engineering output of that, are going to require skill to bring stuff back online. There's another piece with the moats too. If you're optimizing your internal business processes, that's one thing. If you're bringing a product to market against others, the incumbent can go and copy all of the challenger's features. You can use this to strengthen your product by looking at others and incorporating those features as you go. That means an incumbent that already has technical expertise and a certain amount of funds can strengthen their own offering much, much faster. That makes it harder for others to enter the marketplace. As soon as you release your product, someone's just going to copy it.
Dave Sharrock 7:29 Well, I think there's always going to be competition in terms of speed, and there always has been. At the moment we've still got human enterprises, for the most part, using AI as an enabler. So there's still human speed involved in there. What we've talked about previously is social speed over AI or tech speed. So there's still room to get ahead of the competition just by being good at getting things out of the door and knowing where you're going.
Customer Value Beats Feature Velocity (7:59)
Peter Maddison 7:59 Yeah, and I think sometimes, often even, especially if you're too close to the technology space, we come at this from the technology viewpoint and not from the viewpoint of the customer, the customer outcomes, and what the customer is looking for. From a customer's perspective, they may say, "It's lovely that you got that nice new product out there, but I'm kind of happy with the things that I have. I don't really need it to change. It works. So why would I go elsewhere?" If you are already in the market and already serving it, or if you can bring something to market that does what that market wants, there's still room. You see this all over the place within a certain scale. There's still space for selling a spreadsheet that does tax calculations, or figures out payroll, because you don't want to buy a massive solution when your company's really small. Those things still very much exist because they're still valuable products. So I think part of the moat question is: are you actually delivering a valuable product to the customer?
Dave Sharrock 9:19 It feels like that hunt for product market fit that we often talk about. The appreciation of your prospects', your potential customers', unique pain points, and how your solution addresses them in a 10x way, so it's self-evident that if somebody finds it, they're going to keep hold of it and use that product going forward. There are still many of those problems. All of us experience them every day when we're using the technology around us, trying to get things done. Nobody is sitting in some utopia going, "Actually, I don't need anything to change around me. I have exactly the right amount of time for my household chores or whatever it is I have to take care of. It's just this perfect balance." It's just not the case.
Peter Maddison 10:08 Yeah. There was a case this morning. I'm helping out a startup and moving some email accounts around, which is something I wish I had somebody to delegate to. I could have gone and found an agent, made sure it was hooked up in the right profiles to talk to the right pieces, and automated all the different parts of the process. But ultimately it was easier to just go recreate three accounts. Theoretically I could have done that, but I don't anticipate having to do it again for a long time. So the value in me spending the time figuring out how to get everything talking to everything is somewhat limited. Agents, in theory, can go figure out how to automate and make the connections themselves, but they still need the authentication, the passwords, the access, and all the other pieces that go with that. I could have worked through it and done all the bits I needed to. It was just easier to go through it myself. So there's that piece of it too.
Why Traditional Moat Tools Fall Short (11:14)
Peter Maddison 11:14 There are bits around stickiness, and I think it all does come down to value. Does the customer feel like they're getting value from the product they currently have?
Dave Sharrock 11:25 I came from a recent exercise we've done. We've got this new product around getting people to a demonstrable prototype, or even a production-ready product, within five days. From idea to MVP, or to production. As we went through the first few sessions, one of the things that really struck me is that the tools we have to identify moats are very poor. They're generic. Look at Porter's Five Forces, which is a great way of looking at traditional companies, but it really struggles to accommodate where we are in today's digitally connected world, where so much of that power is already priced in. Then the other tools you can look at start talking about network effects, monopolies, regulatory protections, and things like that. The problem is those are global-scale effects. If I'm Uber, I know exactly where I want to get to. I want to become the de facto way of moving around a city anywhere in the world, right? There are massive network effects I can get with that. I need to make sure my app works wherever somebody lands with their phone, and that I have certain things in place. But most of us, if we're involved in a startup, are involved in one with much smaller scale, scope, or ambition, at least at the outset. So finding a moat becomes quite difficult. I think you end up in a place where either you bring something along, like a network of potential users that... wow. Yeah, this is the whole thing with daylight savings. The reason I said "wow" is the light just turned on, and now I can see you fully lit up instead of the haunting shadow of your former self, as it were. So, getting back to the moat conversation, what I was beginning to see is that sometimes you've just got to cast something out there. And the way you do that is either you have access to a pool of users that gives you a big leap forward in some way. I think of Facebook in its early years, when the very first setting was just one university community, proving things out in a small group before going to the next level of scale. Or, on the other side, you get the customer problem understood so well that you give a solution that is as good as it can get for a very narrow niche of potential customers.
Switching Costs And Practical Takeaways (14:02)
Peter Maddison 14:02 Yes. And I think that's it. If you're able to deliver your product profitably to the market you're selling it to, that means the customers want it and see the value in it. That kind of is the moat. Then it becomes a case of: is the customer staying with you because it reflects the value they feel they're getting, or are they staying because it's difficult for them to leave? How hard is it for them to get off your platform onto another one? There's that piece of it.
Dave Sharrock 14:32 Yeah. What I like about this realization is that, in my experience, a lot of innovation teams trying to find the next product, whether it's a startup or internal innovation in an organization, have gone through the exercise of talking to potential customers. They try to understand where they're coming from and what product market fit might look like by understanding both sides of that equation. I've seen very few organizations take that and live and breathe it. And now it becomes more and more vital. Something that was a good idea we could nod our heads to and appreciate is rapidly becoming a no-brainer. If you don't do it, you're not anywhere. If you do it, you'll be able to identify something that potentially has the ingredients of a sustainable, scalable business.
Peter Maddison 15:33 And the ability to experiment with that, bring it to market faster, understand what the customers need, and learn from them. All of those things become so much more critical, but also much easier to do, because I can build it, so at least I can show something. I still need the system and the mechanisms to allow all of that to happen, though, and I'm seeing that not all organizations have them.
Dave Sharrock 15:58 How do we summarize this? What are the three things we might draw out?
Peter Maddison 16:03 In terms of moats, we all know that AI makes the "hey, we're building the new feature" moat a lot less relevant. But customer value becomes critical. Understanding what valuable thing you deliver, and what your customer values about the service you provide, is a key piece of understanding what you're doing, right? We're building for the outcomes, for what we can create. And there's still the market from that perspective. There's a lot of doom and gloom in the tech industry about how you could just point an AI at somebody and copy what they're doing. But just because you've copied what they're doing doesn't mean the customers of company A, the one that got copied, are going to move to company B. There has to be some value, some reason why they'd want to move, even if it's incredibly simple to move from A to B. Why would they?
Dave Sharrock 17:06 We all have subscriptions to products that are suboptimal, but they do what they need to do, and we just keep staying with them through a lack of time to make the change, right? It's not painful enough.
Peter Maddison 17:18 Yeah, it's not painful enough. If it works and generally does what I need, then sure, I can move, but then I've got to learn this other product. How have they done that? And if basically what they've done is point an AI at it and copy it, why would I bother? I don't really want the same thing all over again. There's got to be some reason, right? So when we're talking about moats, I think that's one of the most critical ones still. It sounds great in theory, but the reality is that people are lazy.
Dave Sharrock 17:52 Yeah. I think it's a fragile moat. We're recognizing it's a moat with a small m, not a moat with a capital M. It's not something you can be confident will offer protection over time. However, I also think there are many examples of suboptimal products in the marketplace, and really niche products that we use because they're important to us.
Peter Maddison 18:22 Yes. They're important, the brand is important to us, or it does exactly what we need.
Dave Sharrock 18:27 The experience is just what's necessary, right?
Peter Maddison 18:30 Exactly. And yes, I could go point an AI at it and have the AI build it, and use the credits for that, but then I've got to look after it, maintain it, update it, and deal with all the operational crap that goes with that. I just don't have the energy for that. I've got enough other things to build and maintain and look after. That's just not where the value is for me.
Closing And Next Topic Tease (18:53)
Dave Sharrock 18:53 Just as we wrap up, I want to draw attention to the first part of our conversation, because I'd say that was an unexpected finding. It's the realization that the internal processes that have made us successful to this point can be made truly unique to ourselves. We can build the automation, the workflows, everything else. We build the tools instead of having to rely on what other organizations have. So I think that one is a bit of a secret takeaway from the conversation we've had.
Peter Maddison 19:27 Yeah, I think that's a good point. It is possible to build these things, and the tools themselves, and the harness wrapped around them, are so much more capable than they were before. But that will lead us into the workslop conversation next time.
Dave Sharrock 19:45 Sounds like we have a topic for our next discussion, then.
Peter Maddison 19:48 Well, with that, I'd like to say thank you, Dave, and we'll wrap it up there for the time being. Look forward to next time.
Dave Sharrock 19:55 Always a pleasure. Thanks again.
Peter Maddison 19:57 You've been listening to Definitely Maybe Agile, the podcast where your hosts Peter Maddison and Dave Sharrock focus on the art and science of digital, agile, and DevOps at scale.



