I have used the terms digital and data technology interchangeably. And I have described them as things made of bits and bytes. Of ones and zeros. Instead of physical products and services.
And I have spoken quite a bit about how things made of bits and bytes have different economics. They have powerful but also dangerous economics. The scalability and zero marginal production costs of digital goods are amazing. The ease of bundling can be powerful. But the competitive barriers can also be very low and most software ends up being free.
But, in truth, I was mostly talking about software – and not about all digital and data technology. In particular, I was not talking about machine learning and AI, which increasingly appear to have different economics than software engineering. The efficiency and ease of growth and innovation may not exist. You need a lot more people involved tagging and cleaning the data. So the economics look more like a mix of software and services.
Overall, AI and ML look a lot messier, slower and more difficult than standard software businesses.
And this raises two questions that I have been struggling with:
- How are the economics of AI and prediction different than software?
- Can AI and prediction be a competitive advantage?
I’ll give you my working answer in this article. But first, an analogy for software vs. AI businesses.
Real World AI Is Like Running a Hospital
I did a stint as a turn-around CEO of a hospital in Saudi Arabia. The turn-around part was great fun. It was a lot thinking, diagnosing and then rapidly fixing a lot of problems. It fit my strengths as 90% a thinker and 10% an operator. But after the turn-around, I didn’t much enjoy running the hospital. It was mostly just managing day-to-day operations that didn’t really change. There weren’t any intellectual problems to solve. And I discovered I didn’t like the nature of the hospital business.
Because as CEO, I didn’t have much control. I had 400-500 staff of all types of specialties and skills. Everyone from surgeons and radiology technicians to cafeteria workers and security guards. And I really didn’t have any choice about staffing levels as a secondary care general hospital. You had to have all these people. And you had to have certain departments (pediatrics, ER, OR, ICU, etc.). Which had to be filled with expensive machines. Everything from x-ray machines to blood draw stations. Again, I didn’t have much choice in any of this.
Because as a hospital, you have to handle whatever comes in the door from the external world. You have to deal with all the health issues in the community, which is a complicated reality. If a car crashes nearby, you need to have surgeons ready. If someone with psychiatric issues walks into the ER at 2am, you had to have psychiatrists and social workers ready (or on call). If a pregnant woman randomly walks in and gives birth (this happens), you have to be ready. And if she also has heart problems while in labor, you have to have cardiologists and radiologists ready as well. Oh, and if there is a viral pandemic, you have to have ventilators and protective equipment ready for that.
Running a hospital basically means you have to solve whatever problems nature, society, and medicine might throw at you. You are on the receiving end of everything. So you have to build an operation that can handle a sea of complexity and demands that you don’t control. You are on the receiving end of everything. So you actually have very little control about how you build and manage your operation.
Last year, Elon Musk talked about his cars. And he said it basically required them to solve the problem of “real world AI“. It wasn’t just about driving. It was about having AI that can deal with anything it might encounter in the real world. Trash on the road. Lightning. Pedestrians. Car crashes. And that is why he is now putting this same real world AI into things like robots, trucks and other machines.
That is like a hospital. You have to build it for everything it might encounter. That is messy. And inefficient. And can be expensive.
Now, compare that to running a shopping mall.
Software Is Like Running a Shopping Mall
I also once worked on a high end shopping mall. It had Saks Fifth Avenue and Marks & Spencer as anchor stores. It had tons of brand stores like Coach and Gucci. There was a food court selling everything from pretzels to kebabs. There were a few Starbucks, of course. And there was a Four Seasons hotel upstairs.
And when the doors opened every morning, we knew what was coming. Shoppers were coming into an environment we had designed and created. We decided what was to be sold and what wasn’t. Which stores. Which food outlets. We choose the lighting. We selected the inventory for some stores. We even decided when we opened and closed. And we could even remove people who weren’t behaving appropriately. It was a service business with lots of real estate, just like a hospital.
But in this case, we designed everything. We were only offering what we chose to offer. We were only solving problems we chose to solve. And we had chosen to solve a narrow niche of business problems. We did not try to solve problems in human biology. Or in society. Or in nature.
This is similar to software engineering. Like shopping malls, software is designed to solve a specific business purpose. And a handful of use cases. It is made of components that other engineers have made. It can be tested and debugged. Creating software is a lot like developing a shopping mall, but with much better economics. It really is a different thing than real world AI, which is like running hospitals.
Here’s how I think about it.
- Software is like doing engineering. It’s a system created by humans that is fairly orderly. So lots of designing and building things within business systems.
- Machine learning and AI are more like doing science – with lots investigating and testing of a much more complicated, external world.
You can actually put AI and prediction capabilities into three use cases.
- You can put AI (i.e., predictive analytics) into engineered software that you have created. Like video games.
- You can focus AI on real world AI. Like Tesla and autonomous vehicles navigating the real world.
- You can sort of split the question.
- You can put AI in delivery robots that only operate within a business park, where you have designed the roads for them.
- Or into TikTok, where it is a designed system but the content can have lots of random real world stuff in it.
- Or into search engines that have to predict the right answer to any question anyone might type in. It might be good at common questions like what is the weather today. Or what is the current time or exchange rate. But things get very difficult when people ask obscure questions like who was the most famous nun in the south of France in the 1200AD
It’s all sort of about the long-tail, which is the greatest benefit and curse of AI. TikTok does well at presenting each individual niche taste with long-tail niche content. It’s why it’s so addictive. Google has maintained its search dominance because nobody can match its quality in long-tail search queries.
But it is also the big problem because the costs of effectively dealing with the long-tail is changing the economics of digital. We are not seeing the economics we see in traditional software. It is like going from the shopping mall to the hospital in terms of operations and costs.
Ok. Let me get back to the key question for today.
Question: Can Prediction and AI Be a Competitive Advantage?
For those that have read my Moats and Marathons books, you’ll notice that I put Machine Learning on quite a few levels of this graphic.
It’s one of the new technologies (in red) that is impacting various levels.
It is definitely part of the Digital Operating Basics. It’s going into everything. But those are not a source of competitive strength. They are just operating requirements.
I listed it under Digital Marathons as a potential source of operating advantages. Machine learning / AI is one of 5 important new dimensions of operations you may have to compete on long term. You can see how this would be very important in short videos (i.e., TikTok), inventory (i.e., Alibaba) and in making loan decisions (i.e., Ant Financial).
And it can also definitely create a barrier to entry. Cheap and fast prediction is a core capability that takes time, money and effort to create. That makes it a barrier to entry. However, companies like AWS and Snowflake are making it readily available to companies as a service. So the biggest barrier to entry is going to be the data required and/or the talent required.
And I don’t have it listed as a competitive advantage yet.
My Answer: Prediction and AI Don’t Create Competitive Advantages (Yet)
So here’s my working answer:
- Machine learning and prediction are 100% required operating capabilities. Most companies need to be building this.
- Machine learning and prediction can in rare cases become digital marathons and long-term operating advantages.
- Machine learning and prediction has a barrier to entry, but this is being disrupted. In some cases this will persist based on privileged access to the required data.
- There is no competitive advantage. If there is, it is likely coming from data or network effects.
That’s my answer.
- I think machine learning can definitely be a digital marathon.
- But I’m not doubtful about a competitive advantage. I’ve watching closely for this.
Technology creates tremendous value. But this is almost always passed on to the customer. It rarely gets captured by the company. That usually requires a new business model.
Well, what about data as a competitive advantage?
Final Point: Data is Mostly an Input. Data Network Effects and Data Advantages Mostly Don’t Exist.
An idea that has been floating around forever is the “data competitive advantage“. Or the “data flywheel“. The argument goes like this.
- The more people that watch Netflix, the more data it gathers about what viewers like.
- Netflix then tailors the content to the individual and to the general audience, which improves the service.
- The improved service attracts more customers from competitors, which generates even more data and further increases the service. And so on.
This often called a data network effect. Or a data flywheel. I think this is mostly wrong.
I agree with points 1-2. You can use data to improve services. That’s the digital operating basics. This can be used, in particular, for personalization and improvements in the customer experience, which is critical. My Amazon and Netflix pages are personalized to me. New products and services are continually being launched based on this data. And that’s great and makes the service better. Good.
But Point 3 is a big assumption. And I think it rarely happens. Just because Netflix is improving my service does not mean that YouTube and Amazon Prime are not. Data-driven improvements in services are rarely located in only one company such that everyone shifts. Most digital businesses are personalizing and using data to make their services better. It’s just table stakes for being in digital business.
Data itself is super important and there are cases when it can an advantage. Insurance companies have big historical databases about claims that enable them to better price new policies. We are likely going to see some walled gardens where certain companies will be able to run AI models based on privileged access to data. So we should watch for that. But it will be rare. We could see data advantages (different than data network effects). I am closely watching for these.
That’s where I am on this question. It’s a question I am paying a lot of attention to.
- An Intro to Discount Rates and Cost of Capital for Digital Valuation (Tech Strategy – Daily Lesson / Update)
- Why DCF Sucks for Digital Valuation. (Tech Strategy – Podcast 101)
- An Intro to Digital Valuation (Tech Strategy – Daily Lesson / Update)
From the Concept Library, concepts for this article are:
- Network Effects
- Data Advantages
- Machine learning / AI
- SMILE Marathon: Machine Learning
From the Company Library, companies for this article are:
Photo by mostafa meraji on Unsplash
I write, speak and consult about how to win (and not lose) in digital strategy and transformation.
I am the founder of TechMoat Consulting, a boutique consulting firm that helps retailers, brands, and technology companies exploit digital change to grow faster, innovate better and build digital moats. Get in touch here.
My book series Moats and Marathons is one-of-a-kind framework for building and measuring competitive advantages in digital businesses.