Size Isn't Everything: How AI Will Accelerate Vertical SaaS

In the last five years, SaaS companies have outperformed other venture-backed companies by nearly 5.5%, with the average return on deals between 2000 to 2023 sitting at 23.1% [1]. Horizontal SaaS --- companies that focus on broad use cases applicable to multiple functions across HR, sales, marketing, etc. --- remains the darling of most investors, thanks to enormous market sizes and higher at-bats per home run [2]. But let's face it, chasing every shiny new object isn't always the best strategy. Enter vertical SaaS, the unsung hero that's about to steal the spotlight as AI advances.

Vertical SaaS --- software tailored to the needs of a specific industry, building for a single end market (think construction management, supply chain, HR, nursing home healthcare, among others) --- often gets overlooked by many investors due to its smaller addressable market, fewer outsized outcomes, long sales cycles, and difficult customer acquisition. But that's precisely what makes it a hidden gem. Despite these challenges, the market is ripe for disruption.

A brief history of vertical vs. horizontal

Let's take a look at the history of the software market. The types of solutions companies choose are heavily influenced by the economic cycle with its peaks and troughs. Enterprise budgets move with economic cycles. In times of expansion and free-flowing budgets, teams are able to choose best-in-breed point solutions that were specific to their needs that might be more innovative than incumbent solutions that are slow to change. But in times of downturn and uncertainty, this changes and you might have your IT department breathing down your neck to leverage that seat in Microsoft 365 you haven't been using.

There are many flavors of this when we apply it to vertical vs. horizontal solutions. But, a concrete example to illustrate this would be a legal team using contract management software (vertical) over cobbling together a workable solution via Microsoft 365 suite that the rest of the company is using (horizontal). When your CFO isn't focused on margins and profitability, this would fly. In times of budget cuts and vendor consolidation, it might not unless that solution can demonstrate outsized advantages. In the past, this has been hard to do because output and productivity is hard to measure.

But, that has a high probability of changing as vertical SaaS embraces AI and gives it an edge over horizontal "one-size-fits-all" general purpose solutions that try to do the same.

If we take a look at unicorns (defined as venture-backed startups valued at >$1B and raised a round at post-money of $1B+), there are stark differences that illustrate a general market bias toward horizontal solutions [3]. First, the number of SaaS unicorns is overwhelmingly horizontal, with 339 of the 425 identified companies (80%). Not surprisingly, the average recent valuation ($3.6B horizontal vs. $2.8B vertical) and the total VC funding raised to date ($171B horizontal vs. $46B vertical) skews overwhelmingly in horizontal software's favor.

What is surprising is that the average VC raised to date is almost the same across both horizontal and vertical SaaS. This suggests that while horizontal SaaS boasts a higher number of unicorns and total funding all together, the funding in vertical SaaS might be concentrated in a smaller number of unicorns, given the average total funding is curiously similar. This makes sense since vertical SaaS companies that reach scale typically need to have dominant market share penetration to reach $100M+ ARR. They also need to execute to perfect in order to hit venture scale in a winner-take-all scenario.

But, with the prevalence of AI and vertical software teams building with this in mind, vertical SaaS represents a massive opportunity in the next decade as solutions become AI native. Imagine a construction management software that not only manages projects but also predicts delays, suggests mitigations in real-time, and proposes sub-contractors available in real-time --- a few interesting players are Digs, Parspec, Roofer, Vergo, and Hardware AI. Or, envision medical scribe and coding software that will listen, transcribe, and write medical notes that it will later code accurately for insurance --- Freed and Keebler are already working in this space. Or, think of elderly and disability care software that can enable proactive interventions when patients need more acute care or conversational AI to improve quality of life --- Kangaroo Health, Superbench, and Mon Ami have interesting value props. You get the picture --- that's the power of AI in vertical SaaS.

We're only beginning to scratch the surface of combining generative AI, industry-specific databases, and tailored automated customer workflows that might just turn the tide in favor of vertical SaaS over horizontal.

How AI is changing the equation

While I won't dive into why vertical solutions have endured here (winner-take-all, better capital efficiency, higher net dollar retention, among many others), AI will amplify these advantages:

• More data, more moat: Vertical SaaS can leverage NLP to comprehend domain-specific data. Not only does that mean faster knowledge retrieval and learning, but it also means datasets can be vectorized and parsed with LLMs in better contextualized models. Jerry Chen of Greylock calls this building a "system of intelligence" around data that is unique to a vertical [4]
• Automate, but better: Since vertical SaaS is built tailored to specific use cases within a vertical, its ripe for intelligent workflow automation that just doesn't seem forced --- think better UI/UX and more intuitive suggestions
• Faster product velocity/personalization: Predictive analytics follows a data moat flywheel, which in turn leads to faster product velocity and personalized experiences. Vertical solutions can take data to iterate much quicker and combine it with a true understanding of their end-user needs
• Easier to monetize in the future: Perhaps the most controversial point is that AI can increase the value proposition of vertical SaaS, making it easier to monetize down the road vs. horizontal solutions. Output and productivity gains are easier to measure with vertical solutions that know their end-users and buyers extremely well --- the value/price balance is easier to justify to specific teams than horizontal solutions that claim to do everything for everybody. While I'd hypothesize AI will bring down pricing across the market in traditional seat per pricing, demand tends to go up when pricing decreases and there's massive potential for AI to help historically labor-constrained verticals via efficiency gains

Vertical SaaS solutions are built with specific use cases and ROI for the buyer and end-user in the industry they operate. They already focus on industry-specific data, processes, and workflows. With this context, AI can magnify vertical software's advantages and present new unicorn opportunities in the coming years.

However, there will be lots of competition (and VC funding) as everyone jumps on the AI bandwagon --- ultimately for vertical SaaS unicorns-to-be, it'll all come down to who can execute better whether via smarter GTM or unmatched data moats. It also means the vertical SaaS opportunities out there may be in places that aren't obvious with low NPS, and might even be incredibly specific. The next big thing might be hiding in plain sight.

Notes

[1] Pitchbook, https://pitchbook.com/news/articles/saas-software-vc-exits-early-stage-quantitative-research

[2] For the baseball uninitiated, "at-bats per home run" = (# times a player comes to bat) / (total number of home runs achieved). In simple terms, it's the frequency with which VCs get a significant return (think 10x+) that would be considered a home run. This compares to batting averages, which is (# player hits) / (total at bats).

[3] The base dataset is from Pitchbook. I overlayed SaaS type by classifying companies based on their primary industry focus and whether their solutions cater to specific industries (vertical SaaS) or are applicable across various industries (horizontal SaaS). Don't worry, I leveraged AI and did not classify 425 companies manually :)

[4] Jerry Chen, https://greylock.com/greymatter/the-new-new-moats/