Software is not dead - but your startup might be! 2x AI proof SaaS investment strategies

I am a fan of Harry Stebbings: the British podcaster-turned celebrity interviewer-turned VC. Harry’s content is insightful in part because he does not so much challenge his guests as lets them talk. Take his recent interview with Delian Asparouhov from Founders Fund: 

Harry and Delian covered a range of topics ranging from advice on picking a life partner (it's about skiing) to predictions for Europe’s future (dire) to the outlook for SaaS (also dire). While I do not necessarily perceive Delian to be an thought leader on the first two subjects, it is worth quoting him on the third:

• People like SaaS businesses because the marginal distribution costs are zero

• Stanford kids with AI can replicate your plumber SaaS business in a weekend

So, is SaaS actually on the brink of extinction? 

Read on to learn: 

• What’s wrong with the “AI is going to kill B2B SaaS” thesis

• Why many VCs, including Delian, misunderstand vertical market software 

• 2x blueprints for a future-proof SaaS investment strategy (both involving M&A!)

Build it - and they will come?

Evidently, the views espoused by Delian are far from iconoclastic. In January, Chamath Palihapitiya tweeted that he was launching an incubator that would build custom enterprise software “80% feature complete version at a 90% discount”. 

Is this a viable business model? 

Let’s take a step back. What controllable inputs (i.e. not competition or macro) drive incremental sales of any product - not just software? I count three:

1. Price

2. Product 

3. Distribution

On the first two counts, Delian and Chamath are probably right. The step change in automation capabilities has drastically reduced the cost of building and maintaining software. 

But can a combination of offshoring + AI significantly drive down the cost of distribution? 

For horizontal software, yes.  

For vertical software, not sure.   

Why do so many VCs hold dystopian views about SaaS?

For me, “SaaS” means fundamentally two different industries. One that VCs know well and other that they do not. The power law dictates that VCs chase outliers with big TAMs and as such, are predisposed towards B2C applications or horizontal use cases. Business process automation and sales automation tools are two good examples. Neither requires a tremendous amount of industry specific customisation.

Vertical market software is different. Many sell into tiny TAMs that are both small and static. Consider OMDA, a hospital VMS rollup from Northern Europe. The TAM is only $700M (source). 

Often born out of custom builds, VMS businesses evolve with their customers. You cannot automate decades of relationships.  

Such businesses do not appeal to VCs - and vice versa. 

That said, VCs ignore “legacy” VMS at their peril. Superficial analysis uncovers TAMs that appear gigantic and therefore suitable for startups - when in reality they consist of hundreds of small TAMs already populated by single industry and/or country focused vendors. Their software may appear rudimentary and clunky. But it works! 

The second obstacle that startups run into is the long sales cycle. Highly regulated industries make for lucrative clients. However, they are a pain to land. AI-first distribution strategy works for procurement departments that are run by robots. Alas, I am not aware of any. Procurement teams tend to be heavy on processes and light on risk-taking. I have experienced firsthand how clients would much rather cooperate with an incumbent vendor, than onboard an entirely new one, even if it offers superior product at a lower price point. 

Unveiling 2 AI-proof SaaS investing strategies

Back to Delian’s plumbing VSaaS analogy, if an outsider can build a new product on a dime, what is stopping the incumbent from doing the same / better? 

Subscribe to Premium to read the rest.