As a finance person working in technology, I have always been greatly intrigued by the idea of technical debt. This is the idea that as developers build technology products the realities of time and resource constraints inevitably lead to trade-offs. Eventually, these trade-offs become problems in their own right which require time and resources to ‘fix’. But I think there is a bigger concept out there that we need to grapple with. Technology itself is a form of debt.
Debt does not have a moral weight
Let me start off my explanation by tackling the moral stigma of debt. For various reasons, in our culture debt is seen as something negative. It creates an obligation which makes people uncomfortable. But in the world of finance, debt is just one way to structure capital, this is divorced from any ‘moral’ connotation. (And let’s just table the conversation about what Finance’s lack of moral judgments means for society.) In business school Corporate Finance 101, they teach that companies can raise capital through debt or through equity. Debt comes with a hard obligation to pay back a fixed amount, but any upside once that debt is paid goes to the equity holder. By contrast, equity comes with no obligations, and no penalties for non-payment. But you have to share the upside among all the equity holders. As I said, debt is just a way to structure capital and allocate risk.
From this perspective, debt has serious appeal. If you have an idea for a company or product or project which costs $100, but will eventually be worth $5,000, using debt means immense upside for you — $5,000 minus $100 of debt payback. By contrast, if you raise equity capital, you will have to share a lot more than $100 from that $5,000 pool.
Tech companies typically avoid debt because starting a company means incredible volatility, which in turn means that companies cannot always meet those regular debt payments. And few people want to lend a start-up money for that reason, making it impractical to even consider for most start-ups.
The point is that when appropriate debt provides ‘leverage’, in the sense of a simple mechanical device that greatly magnifies a small input.
Along these lines, I also think “Technology” is a form of debt. It is a resource that comes at a (mostly) fixed cost, that requires ongoing servicing, but one which can yield considerable upside above that cost. There are ways in which this diverges from a true definition of debt, but it is also close enough that it is worth thinking through technology products in these terms.
As with all forms of debt, there are company-level micro considerations of debt, as well as economy-wide macro considerations.
Company Level Debt
At the micro-economic level, there is a general familiarity with the idea of Technical Debt, but I think there is a broader concept at play. If I want to be extremely precise, I would say there are two forces at work. There is Technical Debt which I described above as short-term trade-offs which eventually incur some cost. But I also think there is a broader subject of Technology Debt, this is the ability to derive operating advantage from having technology at work. Put more broadly, Technical Debt is largely a resource issue with which a company’s CTO or VP of Engineering wrestle. Technology Debt is a capital allocation issue that should involve the CFO and CEO. Part of the reason this is misunderstood is that the returns from technology are so much higher than the costs, and so to misunderstand the nature of the technology that underpins so much of the industry. Every company, not only start-ups, can enjoy the benefits of Technology Debt. Large enterprises should (fingers crossed) be able to derive immense productivity improvements through the appropriate use of technology. Just as cash debt can serve as a ‘lever’ to amplify the returns of a company, technology debt serves as leverage along a different, less obvious, path. And this is true not only of tech companies, but of all companies.
Technology Inflation
However, beyond these micro concerns, Technology Debt also delivers macro-economic benefits to the broader economy. Again, we need to remove the judgment we assign to debt as being bad or evil. Debt helps the economy. Debt lets individuals buy houses, debt lets companies invest over the long-term in productive projects. Economists spend a lot of time studying the impact of debt on the broader economy. A certain level of debt stimulates the economy, but these benefits come in cycles which can sometimes get out of hand. Here’s a good video from a smart money manager that explains the idea well, with entertaining graphics.
I think that Technology Debt has also become an important force in the economy. I imagine most people on this site will not find this idea controversial, but I do think it is an idea that is poorly understood. Just as cash debt comes in cycles, I think Technology Debt is also cyclical. We have yet to fully understand this, as so much of what I refer to as ‘technology’ here is so new. In particular, I think some aspects of technology have provided immense leverage to the economy, and we are yet to see what, if anything, the cost of this will be.
There are really only three miracles of digital, technology:
- Moore’s Law. The idea that computers constantly get faster and cheaper has been around for so long, that we all take it for granted. Now that there are signs that it is starting to waver, I think it is important to assess how much of an economic boost the whole economy gets from it.
- The ability to connect all those cheap computers is another huge dose of Technology Debt. I think of this as the Miracle of TCP/IP, or standardized communications protocols, but it should probably just be called reaping the full benefits of Metcalfe’s Law.
- The fact that we can then make use of all those connected, cheap computers through digital code is the most complicated of the three. We can call this the Miracle of Software or Andreessen’s Hungry Corollary. This may not be a single principle. But a survey of the current landscape of programming languages, environments and data structures makes it clear that we are enjoying an incredible blossoming of creativity that is not common in history.
These ‘miracles’ provide benefit largely without cost, but I think all the other products and services we call ‘Technology’ are bound by the laws of economics.
Now I am not a doom-monger, arguing that all of this Technology Debt will slow down some day and bring the economy crashing down. Instead, I think that we have enjoyed the benefits that these forms of leverage provide, and we have probably already seen cycles in those benefits. It is important to track these trends to understand where investment should flow. Viewed in this light, the transition from mainframe to PC to Mobile and Cloud is probably best viewed as debt cycles. The waning years of each of those earlier cycles were definitely less exciting times.
I also think it is important to understand how this form of debt interacts with cycles of cash debt. We are currently living in a time of very easy money, meaning that there is a lot of cash debt out there at very low interest rates. This tends to stimulate the economy, and it is magnifying the Technology Debt cycle in which we live as well.
These two cycles are currently amplifying each other. Start-ups can get funds to build new technology products because it is easy to raise VC money, which itself is easy to raise because of low-interest rates. Those start-ups create tools that enhance productivity and thus contribute to lower interest rates.
Debts magnify the impact of a downturn in the economic cycle
We have to be careful because just as debt magnifies the positive effects, it can also magnify the negative effects as well. For a great understanding of this, read anything Michael Pettis has written. He writes about cash debt in the global economy, but I think it is worth considering how Technology Debt may unwind on the industry. Just as cash debt cycles lead to inflation, we could very well be seeing some form of ‘inflation’ from Technology Debt as well. He goes on to argue that because debt is a hard obligation in down-phases of economic cycles all the leverage we took on the up-cycle becomes a burden in a non-linear fashion. In the case of technology, I think that will mean many projects underway today, will become drags on the economy when things slow down. For instance, imagine a big company building a new technology platform using technology from a start-up. If that start-up disappears, the big company’s project becomes unworkable and much more expensive for the big company to maintain, which in turns drains resources from other projects. We saw that in a big way in the wind-down of the 1990’s telecom bubble. Telecom operators had to drastically scale back operations because they had overspent on technologies that were incomplete. This is a complex comparison, a further exploration of which will have to wait for a future post.
I have been kicking these ideas around for a while, but was motivated to write about them when I read an interesting post by David Strauss at Pantheon called “All Code is Debt” where he is writing about Technical Debt. I do not know Strauss, nor Pantheon, and I do not agree with everything he says, but he makes an important distinction. He says that most companies think about production code as an asset, but he argues that it is actually a form of debt.
This distinction is important because it goes against the grain of traditional thoughts on the subject. Most companies think about all their technology as an asset. Data centers are held on company’s balance sheets as an asset, part of PP&E (Property, Plant and Equipment). And while the building and physical servers are technically an asset, their value is much higher than what is counted on the books. Consequently, companies are under-accounting for this value on their balance sheets. I think that difference should be defined as a liability which needs to maintained or it becomes worthless.
Data is also a Liability
Marko Karppinen makes a similar argument on his blog about Data. In the same vein as I have been discussing here, he thinks that data is not an asset, but a liability that carries significant obligations (i.e. maintenance, regulation, etc.) which we are again under-counting.
Not so long ago, I did a presentation for a potential client discussing the position of ‘data’ on a company’s balance sheet. It turns out that the bodies that govern accounting standards do not have an advanced view of data. Accountants do not let companies treat their data as an asset, nor as a liability. I think there is an important debate to be had on this subject, because most people reading this probably agree that data has tremendous value. The problem is we are still not quite sure about how to quantify, measure, track and (crucially) audit this.
Who Owns Your Cloud Data?
I think all of this is going to come to a head soon. Someday, the economy will slow down, even the tech economy, and slowdowns mean bankruptcies. Just as we are discovering all kinds of great technology in the up-phase of the cycle, in the down-phase we are going to make some truly unpleasant discoveries. These kinds of things are hard to predict (as I argue here and here). But I think there is one area which is almost certain to become very contentious – who owns your data in the cloud? Someday in the near future, a ‘cloud SaaS’ company is going to go bankrupt. And I can guarantee that this will lead to a big court case about the ownership of data on that bankrupt company’s servers. This is not a new concept, CIOs are aware of it and cite it in their reluctance to move to public cloud providers like Amazon’s AWS.Also, we have already seen a few examples of client data being sorted out by the bankruptcy courts. These kinds of events are going to become more prominent.
I have to wonder how many companies closely read the Terms of Service of their SaaS provider? I do not like to make hard predictions, but about this one I am certain. Someone is going to wake up one morning and realize that their customer database is now under control of a bankruptcy judge, and its ownership will be divided up among creditors. Lawsuits ensue. That process can take longer than many start-ups have been alive. In this view, data and many other forms of ‘technology’ are not your asset, they are a liability, in more ways than one.
Returning to my central tenet, we are very fortunate to live in plentiful times. We should not take the creativity and progress which we enjoy today for granted. Someday, perhaps soon, we may have to pay a bill for that.
