I want to talk about Joseph Schumpeter. He was one of the first economists to work on innovation and entrepreneurship, known for the term “creative destruction”.
I stumbled upon a bit of economics history last year and learned that he was behind much of the modern concepts in business economics (That post may be the best short overview of the subject that I’ve seen). I will attempt to further summarize and extend some of the parts that apply more to software and try to tie some of these concepts to movements such as Lean Startup and the Blue Ocean Strategy.
I mentioned the efficient market hypothesis before and Schumpeter’s ideas kind of extrapolate from that. Remember that the EMH says that it’s nearly impossible to make good profits in a free market when there are competitors with the same knowledge as you. Customers will pit similar sellers against each other and cause prices to drop too low for the industry as a whole. This is what makes the economy “efficient” and allows people not to overpay for things even if they don’t know their true worth. It is, however, what makes business difficult since it’s the enemy of profits.
Schumpeter’s Exclusivity Curve
The article on Schumpeter includes a graph summarizing the profit life-cycle. Here is my (augmented) take on that graph:
The graph presents a timeline for “Excess Profits” from innovation projects. “Excess Profits” here, means profits above what you would expect to get, by just investing your money in stocks and bonds instead of in your business. A level above zero means the project or business is worth doing.
One thing this graph shows is that when choosing R&D projects, it’s important not just to think about how much value you will be creating for your customers, but also, how far along you are on the exclusivity curve, that is, how long you have before competition starts offering something similar and causes profits to become difficult.
The need for some kind of exclusivity to make good profits is the reason you hear about differentiation a lot. It is also why, for example, companies like Netflix are focused on creating a constant flow of exclusive content.
A Chain of Exclusivity
Because the Schumpeter curve predicts that generating enough profits to reinvest in growth is temporary for any particular innovation (and in the world of technology, things move fast, so the period can be short), long term growth depends on always having the next exclusive innovations in the oven. The goal is to release this next exclusivity before competition shows up for the current one, thus creating a kind of recursive chain of growth.
Keeping this chain going is sometimes called “traction”. It’s usually the main task of the product team.
When you are at the beginning of an exclusivity period and there are no inexpensive alternatives to what you sell, it’s a good time to “cross the chasm”, to deploy aggressive marketing adapted to a wide market. You basically want to milk the fat margins on that innovation as much as you can while you also happen to have those margins to fund that effort.
If you get near the end of the curve, near the “red waters” and competitors are approaching while you don’t have the next exclusive innovation ready, you have no choice but to focus on finding it to avoid losing traction.
Exclusivity is about Secret Sauces
To maintain exclusivity, you need a flow of new information and new knowledge that your competitors don’t have easy access to.
This often means avoiding developing too many of your upcoming features based on easy to obtain information, based on public knowledge, based on online knowledge or based on knowledge people learn in school. You also want to avoid focusing on sets of features that competitors already have or seem to be working on. Even when these features create a lot of value, profits will be difficult.
But exclusivity doesn’t always have to mean that, vaguely speaking, you’re the only business with a particular set of features (though it helps). It can also be about how these features combine together in a unique way that is tailored for a particular group of people. The details and the execution can matter a lot. It’s possible for solutions to superficially seem like they have the same features, but one solution is much better tailored for a particular group of users because it is based on intimate knowledge of their workflows with many more of their specific use cases done well. This kind of targeting can be advantageous as long as the sub-market is not too small. Modern software often contains hundreds of thousands of lines of code, which means there are a lot of opportunities to differentiate.
Creating the next link in the chain is about finding new secret sauces
Once a product team knows its users very well, finding the next area to tailor to becomes much easier.
Y-Combinator ‘s playbook (now having fuelled companies worth more than $150 billions) describes the importance of creating a stream of intimate knowledge:
“You want to build a “product improvement engine” in your company. You should talk to your users and watch them use your product, figure out what parts are sub-par, and then make your product better. Then do it again. This cycle should be the number one focus of the company, and it should drive everything else.” — https://playbook.samaltman.com/
To stay in control of more unique industry knowledge than your competitors, you have to build a pipeline of information gathering, of analysis and of integration of expertise into your systems. To do this well, you have to cater to a particular group of users. This is why the Lean Startup methodology puts an emphasis on “focus”. If you try to cater to too many different types of people, it’s difficult to build enough expertise on any of them and your solutions will be generic and non-exclusive and thus likely unprofitable.
When your inflow of expert knowledge is high, it also becomes difficult for your competitors to find something you don’t have and to build their exclusivity. This means their margins get tight, which in turn means they have less resources to play catch up. This vicious circle is one of the reasons (on top of network effects) why markets are often “winner takes all” (…until the winner starts dropping the ball, which happens surprisingly often).
In software, often times, the next link in the chain takes the form of integrating business verticals together. You might be able to tell your users: Yes these features are available separately from other companies but we have the only product that has it all integrated and streamlined into a single platform. We can automate a lot more of your workflow because it’s all in one place.
A popular way to find the next link in the chain is to look at what users are doing with generic products like spreadsheets and see if you can build a specialized and streamlined version of that into your platform.
Moating your castle
I won’t dwell on this since this article is already too long but moats or network effects can sometimes help ward off competition for some time and extend the profitable exclusive period.
I will say that network effects are especially important in software that handles a lot of data. It’s hard to offer integrated experiences if the data is held in disparate systems. Good data coordination technologies like Postgresql (the world’s best database) are an important secret weapon for building exclusive data processing logic and achieving network effects. More on this below.
Software supercharges traditional innovation
The article on Schumpeter mentions the experience curve which is one of the more traditional ways to beat the efficient market hypothesis. Because the EMH is based on propagation of expertise and knowledge, companies can rely on finding or cultivating very knowledgeable and experienced employees who are often hard to find for competitors. This definitely applies in modern companies where you can strive for knowledgeable, experienced people that can be good teachers to your users.
However, because the EMH is rooted in knowledge and information, information technology and software is very well suited to tackle it.
It’s great to have expert, “linchpin” employees, but modern companies can also rely on linchpin software having code with years of accumulated knowledge about the industry embedded in it. While employee knowledge decays, — people forget, they can leave, they can even join competitors, software never forgets and never leaves. The amount of exclusivity in modern companies is often a direct function of how much difficult to copy, specialized expertise is embedded in its software systems.
The fact that software doesn’t decay, means that when you embed sophisticated knowledge in it, you are pushing against the limits of information theory and thermodynamics, fundamental limits of the universe. You can get to levels of details that would be impossible for knowledge contained in a human brain. If the code is not public, this knowledge can be very hard to duplicate. It can take years of information gathering and trial and error. Fully exploiting this is incredibly powerful and may explain why software is “eating the world”.
Apart from code and logic, the other form of knowledge that is supercharged by software is user data. I mentioned that RDBMSs like Postgresql are a secret weapon for data integration. People see relational DBs as a place to store data but that’s not what they’re really about. You can store data in plain files. RDBMSs are data coordination engines. They’re the best way to consistently and safely coordinate, join and manipulate data from multiple sources in parallel while avoiding collisions and corruption. Because RDBMSs are so good at integrating new sources of information, it makes doing vertical integration, building the next exclusivity, much easier. And because loosely connected systems or distributed systems are so brittle and inflexible by comparison (provably so), having more and more data in a single RDMBS creates strong network effects while continuously enabling more and more exclusive logic and processes. It’s a weirdly powerful tool to beat the EMH.