This week’s podcast is an introductory explanation to Rate of Learning (and Adaptation). This is an increasingly important concept in digital strategy.
You can listen to this podcast here, which has the slides and graphics mentioned. Also available at iTunes and Google Podcasts.
Here is the link to the TechMoat Consulting.
Here is the link to the China Tech Tour.
Big tech events from this week:
- Expanded export ban for GPUs to China (here)
- Eureka AI training robots in advanced tasks (here)
- iQIYI signs deal with Thailand’s Tourism Authority (here)
Here is my standard framework for digital competition
–—-
Related articles:
- AutoGPT and Other Tech I Am Super Excited About (Tech Strategy – Podcast 162)
- The Winners and Losers in ChatGPT (Tech Strategy – Daily Article)
- Why ChatGPT and Generative AI Are a Mortal Threat to Disney, Netflix and Most Hollywood Studios (Tech Strategy – Podcast 150)
From the Concept Library, concepts for this article are:
- Rate of Learning
- DOB2: Never Ending Improvements
- SMILE: Rate of Learning and Adaptation
- Learning Curve and Experience Effect
From the Company Library, companies for this article are:
- n/a
Photo by Tim Mossholder on Unsplash
——-transcription below
:
Welcome, welcome everybody. My name is Jeff Towson and this is the Tech Strategy Podcast from Tecmo Consulting. And the topic for today, an introduction to rate of learning. Now today, I’m basically changing the format of these podcasts and it’s gonna be one of two things. It’s either gonna be a deep dive into a specific company or it’s gonna be a strategy lesson, a sort of a… quick podcast, probably 20 minutes on what I think is an important strategy concept, and that will be today, which is rate of learning as a concept. Now I’ll explain at the end why I’m doing this, but basically there’s gonna be faster, probably 20 to 25 minutes, we’re either gonna go deep into a company or we’re gonna go deep into a concept. Today we’re gonna do a digital strategy concept, rate of learning, and that’ll be it. Anyways, at the end I’ll kind of explain why I’m doing this, but you know. Let’s just get right into the concept. Oh, quick disclaimer, nothing in this podcast or my writing a website is investment advice. The numbers of information from me and any guests may be incorrect. The views and opinions expressed may no longer be relevant or accurate. Overall, investing is risky. This is not investment legal or tax advice. Do your own research. And with that, let’s get into the concept. All right, now as always sort of start for, you know. What is the concept that’s important? Well, there’s a concept lesson. So the concept is rate of learning. You can go over to jeffthousand.com to the concept library, you’ll see it there. It’s a big idea. I’m actually gonna focus a lot on this over the next year. That’s the concepts. And the other thing I like to do, two to three minutes on important tech events in the last week. And I’m actually gonna keep it to two to three minutes this time. Number one, Eureka. Eureka AI, NVIDIA has released some videos basically applying AI, large language models, to robotics. And it’s pretty spooky, and I’m gonna put the link in the show notes, watch the video, but it’s kind of awesome and scary. They’re basically applying AI algorithms, reinforcement learning to how robotics functions. And the videos they’ve impressed, they’ve released, are basically showing robots doing things like opening drawers, taking a ball from one hand, throwing it to the other, and catching it. And the one that’s gotten people’s attention is basically a robotic hand with a pen in the hand, and it’s spinning the pen. Basically giving robots complex. Tasks and they are learning to do them and it’s awesome and it’s scary look at the video It’s gonna kind of blow your mind at least it did for me. That’s kind of event number one event number two Also an Nvidia story Nvidia has basically shown in their SEC filings that the US government is Expanding the export controls to China 2022 You know, the US government was doing these export bans, mainly aimed at China. It hit some of their bigger chips, the A100, the H100, which are used for AI. Well, Nvidia had gotten away, sort of gotten around this by releasing newer chips, which are a little lower power that evaded, avoided, were not hit by the ban. the A800, the H800, and that’s what Chinese companies have been buying and using. Well, the export ban appears to be hitting those two now. Yep, pretty gangster move. AI is kind of the big thing. Everyone’s using Nvidia chips, and the ban on what Chinese companies are using to do these just got expanded. Kind of a big deal. Third story, okay, not a big deal, but a cool deal. has announced kind of a partnership with the Tourism Authority of Thailand. So those of you in Thailand, it’s actually pretty cool. The TAT and Hai Chi, which is sort of the biggest video producer of China. It’s a mix of Netflix and YouTube. Yeah, they’re doing a sort of a content deal where they’re going to make content about Thailand in Chinese, which will appeal to Chinese consumers. But they’re also going to make content in Thai language. because iQiyi has been going international in Asia for the last year or two. It’s actually pretty cool. So we’re seeing sort of cool stuff happening in Southeast Asia, Thailand in particular. Anyways, that’s kind of a fun story, not groundbreaking, but I thought it was pretty cool. So those are sort of the three stories for today. Eureka AI, the Nvidia GPU band has been expanded and then joint venture and content or some sort of partnership. and I kept it to three minutes this time. Okay, let’s get into the content for today. Now, rate of learning. For those of you who’ve read my books, rate of learning shows up in literally every chapter. I think it’s a huge deal. I don’t think it’s something that’s been talked about a lot. It’s actually as a strategy concept, rate of learning has been around since the 30s. People wrote about it a lot in the 50s. BCG wrote a lot about it in the 70s. It’s sort of coming back as a big idea because rate of learning, how fast does your organization learn? Your organization learn? how fast do your team members learn, how fast do individual people learn, how fast does your organization learn, is becoming a bigger and bigger deal as that learning is becoming digital. And as we are no longer just talking about humans on Teams learning, we’re talking about digital agents that learn. And a couple months ago, I went to the Huawei Connect Conference in Shanghai, which was all about intelligent transformation of companies. Instead of digitally transforming a company, we are gonna do intelligent transformation where we’re gonna build these fairly stunning intelligence capabilities into businesses. So robotic taxis that can learn on their own and go around on their own and give people rides on their own. Businesses that can sort of… look at markets and learn on their own, that can study behavior and learn on your own. Basically, intelligence is becoming the next level up capability that businesses are gonna have to build. And when you start looking into what it means to build intelligent capabilities in businesses, it’s basically an extension of rate of learning that we’ve been talking about forever but was never this big a deal. It’s becoming a big, big deal. And I can kind of tell you where I’m going in terms of strategy thinking. I’ve spent the last five to seven years focused on the question of how do you build a competitive business and how does that intersect with digital capabilities? And that’s really like all these books I’ve written in the last couple of years, six books now, are all about how do you build and measure competitive advantage in digital businesses. I sit at the intersection of competition and digital. But I’ve kind of finished that. I mean, I’m pretty much done. So I’ve been kind of drifting a little bit in terms of my thinking for the last several months, which I think you may have noticed. I’ve been looking for the next big question to sort of take on. That’s gonna take me years. And I’m pretty sure this is it, which is. What does it mean to have a business with intelligent capabilities in the business? So it’s sort of AI meets intelligence meets strategy. And that’s what I’ve been looking a lot into and I’m putting together a project on this right now, which I think is gonna take me a couple years. Anyways, within this big idea, the simplest version of this idea is how does a business learn? Rate of learning. And in my books, you’ll notice that I brought this up at almost every level. So I’ll give you some examples. So I talked a lot about digital strategy having six levels, tactics, digital operating basics, digital marathons, barriers to entry, competitive advantage, and then sort of competitive fortress top of the pyramid. But I put those into two general buckets. The bottom three were about operating performance. How do you compete in terms of operating performance as a business? And as businesses become more digital, that operating performance equation has been changing. And that’s why I broke it down into tactics, digital operating basics, digital marathons. I’ll put the slide in the show notes. The other level, structural advantages, motes, competitive advantages. And I broke that into three levels. But by breaking strategy into these two different levels, which have six layers overall, we can start to look at how digital tools impact each level. And that’s really what the books are about. And that’s kind of my framework for thinking about this. But within each of these levels, pretty much every one, I pointed to rate of learning as an important idea. And in about half of them, I haven’t figured it out. So I talked about level one tactics. If you’re gonna compete as a digital business, a big part of your operating performance is just tactical stuff. back and forth with your competitors quickly, promotions, marketing, discounts, with your customers, a lot of back and forth, all those tactical moves. Well, when you are learning as an organization in a very data-driven way, that sort of tactical back and forth, it’s happening day by day now. Your competitor drops their prices on Tuesday, you drop your prices Tuesday afternoon. that sort of feedback loop is much faster when you start to go digital. Well, as the company gets smarter and smarter, you’re going to start to learn quickly in that regard. So tactics are becoming a lot more about speed of learning. When we move up to sort of digital operating activities, digital operating basics, you’ll hear me talk all the time about you have to constantly improve. You have to feed, you have to have a data feedback loop. where you’re gathering data from your customers and then you’re turning around that into rapid improvements in the customer experience. Well, that’s a type of learning as well. We move up to the next level, we talk about digital marathons. One of my marathons that I always talk about is rate of learning and adaptation. That are you a business that rapidly learns and adapts to a changing market like fashion? So in certain businesses, how fast you learn and adapt is your key operating parameter that you’re gonna sort of compete on. And there’s a book that came out by BCG where they basically argue that adaptation is the new competitive advantage. And as we can move up to the next level, barriers to entry, competitive advantage. I’ve raised the question. At what point does rate of learning become a competitive advantage? And I’m not sure yet, but I think we’re getting pretty close. And then we start to talk about non-human agents. Instead of having humans in your workforce, you have non-human agents in your workforce that are doing things, digital agents. And then intelligence capabilities are almost at the point where if you have a really powerful intelligence capability. a foundation model, Pangu, GPT, Bard, that you are integrating into your capabilities as a business, at a certain point, that’s gonna become a competitive advantage against those that don’t have this. So we can sort of dial rate of learning in at tactics, digital operating basics, digital marathon, and probably a competitive advantage. We’re getting there. So that’s kind of how I’ve been thinking about this. But if you read my books, you’ll kind of know that I, outside of a digital marathon, I sort of just raised the question, but I didn’t answer it. I wasn’t sure at what point this sort of rate of learning was gonna become a real serious, sustainable competitive advantage, but I think we’re getting close right now. Okay, that’s kind of the context for rate of learning as an idea. Let me give you a little bit, that’s about 10 minutes, I’ll do about 10 minutes more and that’ll be it. But let me give you the background for what this is as a concept. I think that’s where we are today. This is kind of the history. Now, rate of learning is basically the idea, it’s almost like a cumulative experience advantage. The more you do something as an organization, the better you get at it. That’s the idea, rate of learning. And it’s usually viewed cumulatively, so. we would have, let’s say, an X axis and a Y axis. On the X axis, we might have cumulative production of an aircraft assembly plant. How many planes have been built? And on the Y axis, we would have some measure of performance. Now, the simplest one would just be cost per plane. So as we get more and more experience as a factory, the more planes we produce over time cumulatively, the cost per plane comes down. Now you can actually change the types of metrics. You could say instead of, you know, cost dropping as the performance metric, we could look at speed, how fast are they built, what is the quality of the planes that are created as opposed to are they cheaper. And you can also look at sort of planes produced, you can look at the metrics differently, but generally on the X axis, we’re talking about some version of cumulative experience. And on the Y axis, we’re talking about some metric relating to performance. And the first versions of this, which came out looking at, you know, the assembly of bombers and airplanes in World War II, or looking at, you know, Henry Ford’s production of Model T cars, the first version of this was cumulative production, factory production versus cost per plane. And it would, you know, decrease in various ways. Now there’s a couple quirks to this. Number one, We didn’t see this in production where there was a lot of machinery. It was mostly, so it wouldn’t be manufacturing the parts of the plane, which was mostly machines. It would be when you assemble the plane because that’s when you have a lot of people. Because it’s the people that get smarter the more they do things. Individually, they get smarter, they get smarter as a team, they get smarter as a factory. So you need sort of a significant labor component. to see the cost reductions. The other thing is the shape of the curve can change. Now it could go down gradually for a while with more production cumulatively, and then it could flatline. Or it could go sort of static, or it could be a sharp drop. Now the curves that everyone got excited about in the 60s, and this is when Bruce Henderson of BCG gets involved, was when you saw a power law. when you saw sort of an exponential decrease. So, you know, the more we would do of something like assemble airplanes, we would see a consistent decrease in the cost, 10% every year, 10%, every year, 10%. And people thought that was pretty powerful, which it is. And so there’s a lot of writing about rate of learning in the 70s when it started to be called the experience effect or the experience curve. the more experience you have, usually people were looking at cost reductions over time and people got a bit too excited and it wasn’t as powerful as everyone thinks, but it can be in certain areas like memory chips or factories. But there’s sort of two caveats to that. Like number one, it has to be in an area with usually it has to be with a significant labor component or you don’t see it as much. The other would be you do lose flexibility when you move down this path. As you specialize more and more, with cumulative experience in making something like airplanes, you do extract those cost savings, but you also, you specialize all of your processes and your people and your training to do one thing. So you lose a lot of flexibility and agility to go from say, let’s make… planes, now let’s go make cars. Well, you’ve hyper specialized to make planes to extract this sort of efficiency. You’ve lost the flexibility to jump between cars and planes and bicycles and motorcycles. So you are kind of placing a bet on a certain type of technology. So there’s some trade-offs there, but generally speaking, we call this type one rate of learning, where we basically say, the more we do of something, we get cheaper at it. We get more efficient at it. That would be type one. Type two would be, so let’s say type one is Steve Jobs, I’m sorry, is Henry Ford type learning. Type two would be Steve Jobs type learning, which is, okay, we’re making iPods and we’re getting more and more efficient at it as our cumulative production increases, we’re doing them cheaper and cheaper, which the iPod was doing. But then we make a jump from iPod to iPhone. Well, that’s a different type of learning. The idea that we can see the market and be creative and innovative and realize that the iPod is a stepping stone to the iPhone. There’s sort of path dependent development. Okay, so now we’re making iPhones, that was a jump. And now we start doing type one rate of learning where we’re doing it cheaper and cheaper with production. Fine, but now we make the jump to the iPad. So this type of learning where we’re seeing these jumps from product type to product type, that’s often called type two rate of learning. I call it Steve Jobs versus Henry Ford. It’s a different process. It’s not all about making things cheaper, it’s about making things better and understanding your market and being tightly sort of attuned to what the market wants and then adapting. So when we talk about rate of learning, we will often say rate of learning and adaptation. In the first type, it was a lot about rate of learning and getting cheaper. In the second type, it’s about learning and then adapting to the market quickly and making these jumps. Henry Ford versus Steve Jobs. Now, type three, and none of this is my thinking. This is like BCG is kind of the biggest thinker in rate of learning historically and even today. Type three would be algorithmic learning. Now, instead of learning over months and years as we produce aircraft, or instead of learning to make jumps, which are sort of larger strategic decisions, now we’re talking about learning on very short time scales, seconds. If something happens in the market, our e-commerce website adapts immediately. It learns from what’s going on in the market. It sees that Amazon has dropped the price on these items, but our e-commerce website sees that and immediately learns, makes a decision, and adapts. And our prices come down. Or if we’re making digital marketing expenditures and we’re putting up ads on Google and Facebook and we see how well they’re doing or not, it’s learning from that in seconds, making decisions and adapting. So it’s rate of learning, but it’s sort of algorithmic rate of learning that is faster than any human can do it. That would be sort of type three. Well, we’re starting to see all three types play out in organizations. And when you put all these together, that’s when you start talking about things like intelligence as a capability in an organization. That we have built systems of people plus digital tools plus AI tools. that are really sort of intelligent in their own way. And now maybe in some cases, it’s just purely the algorithm that is doing it. But more often than not, it’s a combination of humans and algorithms and systems we have built that sort of manifest as a real time intelligence. And the intelligence can play out over years and it can play out in milliseconds. And if one company has built this, and another company doesn’t have any of this, it can be pretty startling for them. And anyways, that would be an introduction to rate of learning and adaptation as a concept, as a digital strategy concept, which was a bit arcane up until recently, and now it’s becoming very, very important in the last couple of years. So the way to think about it is rate of learning. As we do something more and more as… individuals, as teams, as an organization, as a business unit, we are going to see certain performance metrics improve and change relative to competitors without as much experience. The simplest version of that would just be cumulative production results in cost decreases per unit. Henry Ford. Another version of that might be with more experience and cumulative sort of talent and experience, we’re going to be able to make market-facing adaptations quicker. That’s Steve Jobs. And now we’re looking at algorithmic rate of learning, which is kind of a new thing. And the whole thing is teeing up this idea of what does it mean to have an intelligent business where you’ve built robust intelligence capabilities within an organization that can rapidly learn and adapt quickly. I mean, that’s kind of, when I think about intelligent transformation, AI transformation, as opposed to say digital transformation. That’s where I think it’s going. And that’s kind of what I’m focused a lot on right now. Anyways, that’s it. But I think that’s a pretty solid introduction. If you wanna learn more, go to the concept library, just look for rate of learning, you’ll see it there. Some books by BCG about the adaptation advantage. There’s a lot of old papers if you’re really into it. You can go back to the 70s and read about. experience curves, the experience effect, things like that. But anyways, it’s a big concept. We’ll call this sort of an introduction to it. And that is it for the content for today. And I’m right in at about 23 minutes. So yeah, I think this new format is going to work. I’ll cover a concept that I think is important, or we’ll do a deep dive on the strategy of a company. I am going to shift the company mix to about. 50% China, Asia, and 50% US. And let me know what you think going forward. Any specific companies I should focus on, let me know. Any concepts you wanna talk about more, let me know. And we’ll just sort of keep doing these, probably much shorter, sort of quick surgical dives on either a company or a concept. That’ll be it. As for me, just a normal week. Meeting my parents tomorrow. They’re flying to Asia, so we’re gonna tour around a little bit. That’ll be fun. They’ve been here before, so they kind of, you know, I’ve taken them all over. We’ve been to Japan and China and Thailand. So they kind of know the area pretty well. But that’ll be a, yeah, that’ll be a nice week. So I’m looking forward to that. Anyways, that is it for me. I hope everyone is doing well and I’ll talk to you next week. Bye bye.
——–
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.
Note: This content (articles, podcasts, website info) is not investment advice. The information and opinions from me and any guests may be incorrect. The numbers and information may be wrong. The views expressed may no longer be relevant or accurate. Investing is risky. Do your own research.
