It is 2023, and we are not feeling bullish on the future of SaaS businesses.
The recent mainstream conversation about generative AI advancements such as ChatGPT got us thinking about some of the broader implications for the software industry. When we think that with the power of AI, a software engineer can achieve superhuman productivity levels today, the first thing that comes to mind is how AI will disrupt most SaaS businesses over the next decade.
Building a SaaS business is expensive for two main reasons.
Experimentation. First, you need to spend a considerable amount of money doing user research until you come up with a solution that is good enough to be purchased by customers.
Customer acquisition. Then you need to scale as fast as possible and capture as much market share as possible, spending a lot of cash on digital marketing.
The value of some solutions is fundamentally coupled with the number of users on the platform. For example, any kind of market matching platform becomes more valuable as the number of users grows. In such cases, capturing enough market share becomes a barrier to entry for competitors and protects the business's long-term success.
Other SaaS solutions, such as payment solutions, also present obstacles to competition due to the expenses associated with fraud prevention and compliance with central banks and other financial markets regulators such as the SEC or the ESMA.
Most SaaS solutions don't present these barriers to entry for competitors. Think about HR, accounting, project management or marketing solutions. In most cases, the number of users on the platform is entirely irrelevant to the customers (beyond the service provider's financial stability).
The principal value proposition of Saas companies is a good user experience and convenience (e.g. always available, automated updates, no need to install). One of the biggest problems is that patents do not protect most SaaS solutions. Individual algorithms must be patented, but user experience and convenience can not.
The reality is that, in most cases, nothing is protecting SaaS businesses from competition other than the initial upfront investment in experimentation and the customer acquisition costs.
Cloud computing is mature today, and a single developer could easily manage to scale and serve millions of customers. The rise of AI can make individual developers reach superhuman levels of productivity, which means that a savvy developer can copy the user experience from an existing SaaS solution without the upfront research investment and at a fraction of the development costs. The developer would create a product with an almost identical value proposition and minimal operation costs in a matter of weeks. Indeed, the developer would not have the reputation of the original product, but the subscription price could be a fraction of the original and still be profitable.
Another step to consider would be to make the product open source. The development would benefit from free community bug fixes and features, and the market would trust the product more as there would be no fear of vendor lock-in. Ultimately, the developer would only charge a small fee for convenience; SaaS product will "become a commodity product". The small footprint of the team and low operation costs would mean that the business has more agility and capacity to innovate than the large business.
SaaS products will transition from a high-margin to a low-margin model. As new market leaders emerge, the low margins will deter new competitors, setting up the new business for long-term success. The model might is not great incentive for investors, but it will undoubtedly disrupt the existing SaaS businesses and make the the founders wealthy.
First, we know that SaaS business valuations have been using crazy multiples for a long time. The extremely high valuations are not a very good starting point. The first step would be dramatically reducing operating costs and transitioning from high growth into profitability so you can reduce your margins as time goes by. The business's ultimate goal should be to charge for convenience, not a product. It would be best if the company ultimately considered open-sourcing the product.
Another thing that the business should consider is to leverage the data from the existing large numbers of customers and embrace AI as soon as possible. Certain AI features require vast amounts of data and can sometimes be patented. The lack of data and the patents could protect the business against disruption from new market entrants.
What do you think? Is this too far-fetched? It won't happen overnight, but time will tell; we should find out in the next 10-15 years.
Thanks for reading!