The SaaS business model exists to meet the needs of its enterprise customers, specifically:

• To stay productive and efficient (automate workflows, single source data, etc)

• Save money and resources when compared to the increasing cost and complexity of maintaining on-premise software systems.

• Access software on-demand, without the need for extensive installation, maintenance or multi-year contracts

The above assumes that:

• Everyone has access to high speed internet (low latency response times)

• Engineering talent is scarce and expensive

So in the current state, given great internet access and high cost of engineering talent, it doesn’t make much sense to build something yourself when you can sign a contract and have access to the latest and greatest product in a few minutes.

SaaS companies have benefited much from the current state, reflected in their 70–80% gross margins. The tech industry has the highest gross margins when compared to any other industry.

Is this sustainable? Why hasn’t this been competed away? Would it ever make sense for customers to build something themselves? Why aren’t the gross margins competed away like other industries?

Two anecdotes:

I am in the process of winding down an AI start up. The start up was trying to help enterprise product teams analyze and action customer feedback across spoken and written customer feedback (calls, surveys, emails, tickets etc). We existed 1 year before ChatGPT was released, and wound down 1 year after. Before ChatGPT came to market, our target customers had no desire or resources to build AI models themselves to solve their problem. The cost and complexity of NLP talent and NLP models was too high. Our ACV sat around $20K USD. Not great, but a good starting point. Our pipeline was filling up, and we started to convert a deal or two a month. The machine was starting to come to life. After ChatGPT launched, things changed. Very quickly the market was flooded with GPT wrappers. In house development teams could now access the worlds best language model with a few calls. Sure it was slow and expensive, but there was very little complexity and any dev could do it. Almost immediately we struggled to close deals. We were booking meetings, having great demos, but closing became a big problem. As a test, we took 100 open opportunities and ran some pricing promotions to see if the problem was pricing related. The results showed that customers were no longer willing to pay $20K, but would pay $2,600 (anually). Given the high compute costs, 70–80% gross margins were laughable. At that price point, our GTM was completely broken. We called a few prospects and had some frank discussions about price. Three themes emerged: 1) Other tools were priced in the $10–$25 per seat range and were bundling some of our AI capabilities at no additional cost. 2) They could build it internally (or believed they could) at no incremental capex. 3) They didn’t need a polished product as it was an internal tool.

Given our short runway, we decided to pull the plug.

A non-tech example:

I live in rainy Vancouver on the North Shore with lots of tall trees. The exterior of my house needs a yearly clean up (gutters, decks, moss removal etc). I got a few quotes for it, all in the $3–5K range. Although I didn’t want to, I could do it myself and guessed 4 full days. I wouldn’t do as good of a job, but it would be sufficient. I had access to the same tools (power washer, ladder, moss cleaner, etc) and estimated it would take me two weekends. I did it myself. If the quotes came in at $1K, I would not have done it. The point is, at some point, it stops making sense to outsource things even when you don’t want to do it yourself. And this will ensure gross margin is inline with most other industries.

I believe the same it true for SaaS and 2024 will be the first year this becomes clear. What’s changed:

• Access to engineer talent has gone up through increased comfort with remote workers and offshoring.

• Increased productivity of existing engineering talent through no code/low code products

• Enabling non-engineers, non-product managers, non-product designers to become engineers through no code/low code products

The bleed will hit vertical SaaS tools first, especially those who’s product is an internal tool for their customer.

Will SaaS disappear? Definitely not, but I think the 70–80% gross margin days are over.