My partners and I will make our first pilgrimage to SaaStr this week. The tweet at bottom from my colleague Jackie should give you a sense of how SaaStr feels for a SaaS focused VC. It’s like a StarTrek Convention for a Spock impersonator. SaaStr and – twice in one week – being asked “what does AI mean for SaaS?” got me thinking about the “essence” of software.
There is only one other technology in human history that has had the impact of software – the printing press, of course. (okay, maybe electricity too) However the growth of literacy rates that printed books spurred pales in comparison to software adoption, using internet as a proxy:
Source: Ourworldindata, OECD, UNESCO, Scribblrs
Why is this? The marginal cost (and often price) of software is truly approaching zero, putting it within reach of even the poorest of the poor. The marginal cost of a printed book has never been that close to zero, so the price and effort required to learn letters has long been insurmountable for poorer tiers of society. The good news is software itself has now too made the marginal cost of a book zero, which in turn will help drive literacy rates on their final climb towards 100%.
Okay, so software is “eating the world” rapidly, but what does this mean at its core? In short, software is a digital means to drive marginal costs of transactions of any kind – human-to-human, human-to-machine, business-to-human, and so on – to zero. This definition is pretty broad, but purposely so. From punch cards replacing the human “computers” of Hidden Figures to a bot managing customer service interactions for a banking app, it is all part of the same trend. Software can help humans do stuff faster, cheaper and more accurately, and each new wave of technology (eg, now AI) brings software that much closer to surmounting costs of existing transaction methods… and displacing them.
Nerdy, true, but I am using this specific economic framework purposely to lay groundwork for the AI question. I am not excited about AI itself; it is just another tool in the expanding toolkit of technology infrastructure that underlies software. What I am excited about is the specific situations in which it can make software work cheaper, faster and more accurately such that non-software processes can be replaced.
We’ve invested in two examples of this, both in Ann Arbor, MI. Why there? Surprise, surprise not every developer knows how to do AI just because they are a developer (a word of caution to the AI snakeskin salespeople out there… we’re on to you). There is real AI coming out of The University of Michigan. Notion uses AI to help me communicate with my colleagues more efficiently by knowing what and who is important. I use it instead of gmail. Clinc lets banks offer personalized chat interaction with consumers, foregoing frustrating phone trees, waiting and customer service reps’ own shortcomings.
But AI is only the most recent in a long line of infrastructure innovations that drove (and continue to drive) the adoption of software to the zero marginal cost singularity:
Silicon transistor (and Moore’s Law) –> Workstation/mainframe –> PC –> Internet –> SaaS (cloud) –> Mobile –> IoT –> AI
There are also UI innovations I’ve discussed that have a similar effect:
Punchcard –> Keyboard –> Mouse –> Touchscreen –> Voice –> VR
As with AI, when any infrastructure or UI innovation emerges, there is always buzz of exponential promise. However, entrepreneurs and investors need to see through the hype to specific use cases where transaction costs are actually diminished by the new innovation, and therefore adoption is warranted. We believe, for example, we are doing this with our investments in Notion and Clinc. A cool infrastructure or UI shouldn’t be used just because it can be and is not a customer benefit in itself. for this reason, I cringe when I hear startup one-liners like “AI for XX” or “ML for YY”. In many cases, the best innovation implementations will barley be noticed at first as with Google’s use of AI for image recognition, now a powerful consumer benefit in Google Photo that enticed me to switch from Dropbox. Google doesn’t call it AI; the new feature is simply an obvious enabler of search that makes search faster and more accurate.
So at SaaS this year, let’s not get caught up in the hype of technology but see the long arch of software for what it is, an unfinished journey of removing transaction costs in business and all facets of life.
Bonus 1: Internet adoption inflection point with mobile
In looking more closely at the internet adoption curve, it is amazing to see the inflection point to a higher adoption rate in the late 2000s. What caused this? The first major trend of linear growth happened from 1998 to 2007 with the PC boom while the second major trend of linear growth – at a higher rate – happened from 2007 through today with the mobile boom. This is driven by India, China and South America where mobile leapfrogged gaps in electricity, phone and other infrastructure in poor and/or rural society. No doubt access to mobile internet will help drive the final journey of global literacy rates in these places as well.
Bonus 2: VC excited for SaaStr!