The tech industry has been obsessed with commoditization as a market trend since the crash. For most, its simply a chicken littlish "my prices are falling!" But I know a little too much about commoditization from a former life, co-founding a commodity exchange for bandwidth and leading a risk management software company for a time. My company moved out of the telecom market when it crashed into equities, but others remain. While building an exchange, we had to work with energy and other commodity traders to understand their fundamentals and apply them to markets otherwise driven by innovation. Some of the research we did back then can be found at Oncept, including how we identified market characteristics unique to telecom such as temporal and geographic arbitrage. So before I moved up the stack, I learned just enough to be dangerous.
Let me provide some definitions and then apply them to the commoditization of software.
First, a commodity market exists when:
- There is a common definition of the good and standardized contract. This includes its quality or service level and keep in mind there are over 100 gradients of West Texas Intermediate Crude, the most liquid oil contract.
- There is enough liquidity, or volume of trading in the market
- There is no concentration in supply
- There is no concentration in demand
- Pricing is volitile and indexed
Second, the Internet is a force for commoditization. It aggregates liquidity in otherwise fragmented markets, make pricing transparent and reduces transaction costs.
Third, all markets trend towards commoditization and prices decline. A market where prices have a rising trend is in contango, when prices trend to zero the market is in backwardation. Innovation creates new products and bundles that are initially in contango, but backwardation raises its ugly head at increasing speed with competition.
Fourth, when commoditization occurs the market grows in volume. Inflection points usually happen when there is a market failure that identifies new risks. When new risks are factored into models, the new risk management practice enables greater liquidity.
Fifth, there are four outstanding risks to manage:
- Market risk -- price volatility
- Technology risk -- obsolescence
- Operational risk -- execution
- Credit risk -- counterparties and exposure
There is little doubt that MIPS, Mbs, Mbps and Mhz meet the criteria for commoditization. If you are a player in a Datacommodity market, here's some ways you can make money:
- Volume, through realizing economies of scale and speed
- Trading, through arbitrage
- Bundling, through realizing economies of scope and span with an adaptive infrastructure
- Risk Management, through offering your customers bundles that shield them from volatility. Above all, this is managing complexity
Common wisdom in the valley is that when commoditization occurs in your market, you have to move up the stack. This is partially true, but forces some companies to move into services too early or without a competitive advantage. Instead, the first step is to address risks and inefficiencies -- to innovate in marketing, not as in spin, but value chain orchestration and commercial configuration. Embrace commoditization, since it is inevitable. One positive trend is companies already up the stack now take commodity managment as a core competency (e.g. Google's hosting complex) or are addressing it with partners that offer Grid Services.
Now on to software. This week, Jonathan Schwartz commented:
For software to be a commodity, intellectual property is a commodity, and I'm not buying that.
Unfortunately, the vast majority of commodities are intellectual property, not physical goods. They are derivatives. When someone writes an option, the intellectual property is rights afforded in the agreement. Currency, the largest market of a good, is also an information good. But Jonathan's point may be that IP goods are so maliable its hard to see them as fungible.
The software industry was due for commoditization, but something disruptive happened -- the rise of open source. At the time when competition would usually prompt the creation of a standardized commercial agreement, open source agreements like GPL were created. In other words, monopoly positioning commercial competition prevent necessary cooperation to pool risks and grow the market profitably. This could not have happened if the Internet didn't allow open source advocates to find each other in numbers and collaborate, not just in commons-based peer production of goods, but the commercial rights to them.
Now the industry is pooling risk in two areas: technical standards and open source. Standards are not about a common definition of a good for trading, they are interfaces for interop, which reduces transaction and switching costs (reduced lock-in and ability to bundle for customers). Besides the economic efficiencies gained, they enable easy combinations, or bundling, which drives innovation up the stack.
Open source reduces risk for participants of competing in markets where they don't have advantage. They contribute to the pool and benefit. Current open source agreements are awash with free options that could be seperate agreements. Their constraints on commercial use do not serve many markets. A standardized agreement reduces transaction costs and attracts liquidity. Especially when this contract has no competition. I have to wonder how long it will take for existing players to wake up and standardize a commercial agreement. The market has room for both open and commercial agreements and all software business models will trend towards hybrid (e.g. Socialtext has a hybrid open source business model that provides both open source, hosted service and appliance options to customers). I'll take the ASP Geneva Agreement with service schedule five, please.
A commodity market exists for software, primarily in components under open source agreements. In most commercial markets for software, there is too much concentration in supply and lack of standardized agreements for there to be a commodity market. Instead, commercial markets are using open source inputs. Witness the raft of startups building upon the LAMP stack.
But there will always be innovation in software and not all software will be a commodity. The trend is more software being built upon commodity inputs, mostly open source. While software that automates business process have grow into largely mature saturated markets, there is room for innovation. Much of the low hanging fruit is in managing the complexity we have created or orchestrating new elements of the network. But software for people goes largely unserved. The vast majority of employees are not engaged in process, they manage exceptions to process. Each wave of innovation leads to standardization which begets innovation again.
What's interesting is how players like MySQL are fostering a commodity market of open source derivatives while de-commoditizing with commercial services. They manage risk, complexity and quality for customers. As a vendor, they benefit from innovation in a liquid open market to reduce their own technology risk. Their pricing is disruptive and their role in orchestrating the value chain allows them to mitigate market risk.
Embracing commoditization doesn't just have to happen in areas where the software is core infastucture or the market player arose from open source. Its a simple acknowledgement of risk and the foresight to cooperate with erstwhile competitors.