MOATs are dead. Long live brand
It’s a tale as old as business itself. A company becomes large selling a product or service, either creating a new market or dominating an existing one. For years they enjoy the success and profits of their position, and for a time it seems impossible to imagine that market without that brand. Sometimes the name of the brand becomes the shorthand for the market itself.
Blockbuster. Hoover. Intel. Ford. General Electric. Woolworths. Borders Books.
The idea that they will always dominate the market isn’t just fanciful thinking. Smart business people and academics would study them, and eventually call them ‘Competitive Advantages’ or ‘Moats’. These are very good reasons why a new player couldn’t simply come along and scoop up the market.
Here are some examples:
A lot of these make total sense. It is very expensive indeed to make a single car, but probably much cheaper (per car) if you can build a factory that can build 30,000 of them. But in order to get to the position where you can build a factory, you need to have made lots of cars that have sold and people have loved them. It just isn’t that easy to create a new car company. If it was, everyone would be doing it. I totally would be.
And yet, even against this backdrop, titans fell with regularity.
Blockbuster and Hoover are no more, but not because people didn’t want to watch movies, or clean their house anymore. Yes, poor decision making and mismanagement didn’t help, but in both of these examples (and countless others) a shiny new entrant had smashed their way into the market and were now flying (hello Netflix and Dyson).
Countless column inches and business books have been dedicated to these ‘disruptors’ — these Davids that had toppled Goliaths and had now taken their place.
Netflix. Amazon. Facebook. Uber. AirBnB.
All of these had seemingly overcome the myriad moats and competitive advantages and become the new top dog.
The idea is that the incumbent (the big slow company) just doesn’t have the agility, willpower, vision and force of will to drop everything to innovate and avoid destruction. Their hubris is fuelled by ‘our customers won’t leave us, they love us’ or ‘we’re three years ahead with our tech and no one can catch us’. They capitulate because of internal politics and profit targets, and are simply too risk averse. They don’t see the writing on the wall until it’s too late, and then become resigned to the history books.
We collectively enjoy these stories — this is part of what fuels the ’10 hottest start ups in the world right now’ — we want to be the first to spot the next giant killing mega disrupter. Early adopters love getting their hands on products early and riding the wave all the way.
Those of us that work in tech are always primed to be on the hunt for the latest unicorn, be that in the world generally or in our own industry. When the competition comes for you in tech, it can be bloody, so most start and scale ups always have at least one eye on the competition.
Then, over the last two years, something really really weird happened.
OpenAI, the giant killer?
Just over two years ago, OpenAI launched ChatGPT. At the time, it was viewed as near sorcery. It felt like we had taken a giant leap towards AGI, the singularity, skynet.
According to OpenAI, ChatGPT acquired 1 million users just 5 days after launching in November 2022. It became the fastest growing product of all time.
Who were these nerdy researchers, and how on earth on earth did they just magic something this monumental from thin air? Surely this was the beginning of the end of an era. The tech Titans of Silicon Valley, once seeming so invincible and uncatchable, now looked far more vulnerable. How would Google search be able to compete with an AI in your pocket that could not only answer anything, but format the answer so beautifully, and even say it out loud to you?
Then the weirdest thing happened…
Within 18 months, Google, Facebook (Meta), Microsoft and others had launched their own LLMs… and they were pretty good, actually. These giant, slow moving behemoths had actually caught up pretty fast without really breaking a sweat.
Huh? Like a Silicon Valley Roger Bannister Effect, instead of toppling these giants they had simply opened their eyes, and rapidly released their extremely competitive versions within months. This isn’t how it’s supposed to work Google. You’re supposed to be dying slowly but surely, full of confidence and bluster that AI is just a phase.
Turns out, that though the breakthroughs from OpenAI were impressive, their secret sauce really came down to scale. More data = better models. This meant lots of data. And turns out, there was plenty to go around — 30 years of the internet to be precise. Google, Facebook, Microsoft had access to LOADS of it. Everyone had access to the internet.
So OpenAI’s new competitors copied their approach and got pretty good results. There was no outright winner that was better, faster, cheaper than everything else, and no one was uncatchable.
Even more recently DeepSeek showed you don’t need the deep pockets of these companies. This was now an arms race taking place faster than anything any of us had ever seen before.
Fist came the LLMs, then came the new generation of AI companies
These LLMs were often cheap, or even open source. Given their general purpose nature, this meant they could be ‘baked in’ to any software, provide intelligence underneath a user interface.
Given the breadth of knowledge these LLMs have, the number of industries where fast moving, funded start up could start a quest to disrupt, was absolutely massive. The crazy part was that it wasn’t even the traditionally competitive software market likes sales software, or marketing analytics. These LLMs could compete in arenas that took humans years to qualify to know how to do well.
For all of us, it really comes down to this:
If your product or service touches data that is available in any large quantity in either public or easily accessible domain, AI will destroy most (if not all) of your traditional competitive advantages.
This is because AI does a really really good job of consuming all of that data, and churning out very passable average content based on that knowledge. Medicine, Accountancy, Law, Finance, Banking, Marketing, Engineering, Framing, Education — all take humans years to get good at, but AI can get pretty good at them, really fast.
It can even write code and help you build software quicker. Combine these two and any number of start ups will spring up in industries previously seen to be protected by domain expertise, or specialisation, or domain expertise.
Take these new companies:
- A will writing service, powered by ChatGPT (no lawyer required) — https://www.willgpt.com.sg/
- AI powered accounting (no accountant required) — https://trullion.com/
- AI music creation software (no stock music required) — https://suno.com/home
- Voice Generation (no expensive voice actor required) — https://elevenlabs.io/
- Error detection for data (no expensive engineers required) — https://cleanlab.ai/
- Write code (again, no expensive engineers required) — https://codeium.com/
5 years ago, companies that offered a version of these services would be judged to have significant competitive advantages against their competition. Now, powered by LLMs, dozens if not hundreds of companies can be spun up offering largely the same service, often at a fraction of the cost, since they have not amassed the underlying cost of service.
How then can companies in industries ripe for disruption by LLMs and AI stay competitive?
A deeply human understanding of your customers
LLMs and AI can answer a broad range of questions and give some pretty good answers, frameworks, approaches, code and more.
It is however, by definition, an amalgamation of publicly available data, condensed and formatted to your needs. Even with some heavy prompting, the outputs of software built on this is likely to give you something well within the average.
Really successful companies however aren’t satisfied with average. They build products, market and position them with a deep understanding of the world of their customers. This is unique, and untouchable by AI trained on the internet.
Easy Febreze-y
There’s a now legendary story about the launch of Febreze. When the underlying technology was invented, it was a genuine scientific breakthrough. More than simply an air freshener that pumped nice smelling particles into the air, it actually eliminates odours. Initially this was marketed to people who own pets (think very smelly dogs) as a way to stop their homes from being a shock the the nostrils for anyone visiting.
The problem was this was twofold. One, people with pets has become ‘nose blind’ to the smell of their own homes. Two, people don’t feel great about being told their homes stink.
The answer was relatively simple — instead of starting the process of cleaning your home by dowsing everything in Febreeze, the new adverts showed people first cleaning up and then adding a sprtiz of Febreze at the end of the process — a final touch of lovely smell.
What this showed is a deep understanding of the psychology of their potential customers — something AI could not give them, and then allowed them to reposition their product to capitalise on this insight.
The product didn’t change, but the insight into customers and their behaviour did. For your products, this kind of insight might just protect your brand.
And talking of deep insight….
The unsexy, bottom of the iceberg stuff
Surface level features are just that. Fantastic looking on a website, great for beauty parades, fantastic for product marketing collateral.
But these are rarely the reason people really buy products.
Everyone who has ever adopted a new piece of software (particularly when swapping it out for an existing one) will know that those feature represent a tiny fraction of the pain and frustration experienced of actually embedding something into a tech stack and existing workflow.
What most of us are looking for is that we’re buying from a company that understands what we’re trying to achieve and has taken steps to make that possible. That does comes from features, but also the language they use and the way they demonstrate that they understand our pain and our world.
This comes from changing your marketing from talking about yourselves, to talkin gabout your customers. Of course, along the way you can mention things you have put in place to make their lives better, but it should always be grounded in a deep empathy with their situation.
This might mean that integrations, and help getting set up with complex, legacy systems is actually a far bigger selling point in the world of AI that any given flashy feature. For larger businesses in particular, it may not matter that answers to any question can be returned within milliseconds about a customer, if that fancy AI system actually cannot be integrated with your old school customer database.
This stuff, these hard lessons learnt, these bottom of the tech iceberg stuff might actually become your biggest strength when faced with dozens of AI companies who claim they can do 90% of the same stuff you can.
Brand affinity
Another hard lesson might be that your potential market just got a lot smaller. In industries where companies like yours was the only option, there is a new breed of fast moving, VC backed, Silicon Valley based companies run by 25 year olds gobbling up prospects who are in fact looking for the latest shiny thing.
Hopefully your market is big enough that there is still plenty of pie to go around for you. How can you ensure you still get that slice of pie?
Again, it all starts with your customers. What type of business are they? Why do they like working with you? What kind of people are they?
This kind of psychological insight can lead you to something a little bit magic — brand affinity.
In Pulp Fiction, Mia says that people are either Elvis people, or Beatles people. Most of us can take a pretty good guess after spending time with something if they are an Apple or Android user, a Windows fan or a Mac fantatic, a dog or a cat person.
My favourite piece of recent brand affinity marketing was done by Fiat:
The advert tells you absolutely nothing about Fiat cars, other than they don’t come in grey anymore.
If you want a grey car, that’s absolutely fine, you’re just probably not a Fiat Driver. If you are passionate, love food, love love, embrace the full colour of life, then you probably are.
This kind of brand affinity is hard to build, but if you can achieve it, your customers will stay with you long after the latest shiny AI-powered product comes out.
Building an emotional connection with your customers is a competitive advantage that can withstand the onslaught of new AI powered super start ups.
In fact, it may be the only one. Time to get serious about brand.
