bye SaaS

My intention for this essay is high volume. I care about getting shots up more than I do about legibility (at least right now, in this moment, as I write this essay). So do not judge success here as coherence. Success is merely volume. Literal word count. Idea count. Ideally I get to some original ideas as I dig through the noise of beliefs that may or may not be paying rent.

The “leading edge of innovation” over the past few decades, at least in the public forum, was “software businesses.” Tech companies ran the table. Like they were cool. Celebrities wanted to invest in them. CS majors dominated. Ran the table in the public zeitgeist. Social network was a movie that people know the soundtrack to.

I do not think it was always this way. Like in the 1960s this was not where cool lie. It changed. And I think that it will all end.

Like how many companies do you know from the Sequoia portfolio in the 1990s? I think there are probably a single digit count. They all went away. There are waves and bubbles and cycles. Financial engineering. People grab profits but they all come crashing. These are not sustainable things. They are not default great businesses just because they are new. They are not long term businesses just because they are growing quickly.

I am skeptical of software as a service businesses (SaaS). Skeptical not because I do not sometimes like using them and not because I think they are all massive failures by most people’s standards. More just that I do not know if it will be like the thing 100 years from now. They may just not be that big. Like compared to full stack. Compared to owning the end to end service. To guaranteeing outcomes for your customers. SaaS is just infrastructure to help other companies be productive. Like sure you can use infrastructure. Infrastructure tends to be a commodity which trades basically on cost. So like SaaS can be infra but will infra grow that quickly forever? Probably not. So what you end up doing once you realize that you are a commodity is build a big sales force so it becomes more expensive to sell things. Or you try bundling. Like you build a bunch of other commodities thinking that together, together they will be differentiated. It is all sales. AI will make the building of the thing truly a commodity. It already basically is. Like you trade supposedly high margins for total addressable market. I think what you are really trading though is experience. Customer experience of controlling the end to end experience.

People stopped building full stack. It was too scary. VCs said so. They said it wouldn’t work. Too hard to scale. Margin profile didn’t make sense.

And maybe they were right as they used the pattern matching algo they had been trained on from previous cycles.

But how big of businesses are we actually trying to build? How long should they last? What are we actually trying to do here?

We may be reaching the end of the SaaS era. There are all of these productivity tools on the market but none that really make us super productive. There are all these communication tools but none that make us love communication at work. The list goes on. It’s not clear to me that SaaS businesses will be remembered particularly well, if at all. 200 years from now. It is not clear we will be proud of these businesses. It is not obvious this is the right place for us to be spending our energy. Spending not investing. All the CS majors from Stanford building supposedly modern versions of google sheets. Is that really the outcome? Putting fancy illustrations on our landing pages. Making it seem like we are building tools for thought and changing the world. Is this really the future? Is this “based?” Is this first principled thinking? Why work so hard? Is anyone even going to remember this stuff? Who cares?

I would imagine the future looks different. Or at least could. Or hopefully should.

I imagine we end up with more architecture style firms. 20 people. Like a research lab. With high standards. And high scale. And high margins. A new culture of building where less is truly more. Where you can be ruthless about prioritizing the things that matter (like profit) and not be viewed as culturally inept.

And then we need more assembly lines. To implement the ideas. Using end to end digital coordination tools that run the whole machine. To direct the people more effectively. More coordinators and librarians. That is the point of tech. To take the thing us humans keep getting stuck on—interacting with other humans ineffectively.

And more ambitious visions. More people who want to own the thing end to end. Enough slicing of the pie. More building pies.