Replacing SaaS
Smart operators are going all-in on Vertically Owned Software
Companies that build software in-house will win their market.
It’s always been true that proprietary software gives you an incredible edge. But, prior to 2025, it was cost prohibitive to build.
Instead, CEOs purchased horizontal SaaS (Salesforce, Hubspot) and more recently vertical SaaS.
Both of these are great but have two big downsides:
All of your competitors are also using them (where is the edge?!)
They never work quite right for your specific operating model (read: inefficiencies)
I’m leaving “cost” out of the downsides because any alternative will cost you money, too. But now, the build vs. buy scales have shifted dramatically toward build.
To demonstrate how in-house software can give your company an “edge” — consider how things played out for a tech mortgage company I helped develop:
Everyone in the industry used the same lead aggregator and CRM. Landing a customer depended on being the first lender to dial their number from the call center. Building a dialer in-house gave us a several second lead on the competition, allowing us to consistently win the first call and ultimately making our marketing spend WAY more efficient.
Replacing SaaS
Horizontal SaaS tools need to be everything to everybody.
Vertical SaaS tools need to be everything to a specific slice of everybody.
This inevitably means that lots of the features in a SaaS product will go unused by the typical user.
It also means that many boilerplate features are duplicated across every single SaaS.
Here are a two examples:
Every SaaS implements functionality like authentication, user accounts, teams, and billing
Every SaaS maintains a robust “integrations” library – even though any one client will only use 2-3 of them
Thankfully, companies like supabase have made the boilerplate features very simple to build in-house.
Chatgpt and Cursor have made building integrations a breeze. For example — I wrote a medium-weight slack integration in a few hours that would have taken me a week previously.
This means that companies building in house can focus on the specific features that will allow them to differentiate.
Why now
The cost of building software is plummeting rapidly. I’d estimate that it may be 80% cheaper today to build and maintain software compared to two years ago. Additionally, the last decade has been defined by two major trends:
The API-ification of “lower-level” functionality (ex: Twilio for calls and texts). There is an API for nearly everything now.
The rise of open source software. There are now free packages to handle much of what used to be coded in house.
The chance that your software will need to solve something “novel” is nearly zero in most industries.
The pairing of APIs and open source means that the LLM is mainly focused on stitching different APIs together. This changes the developer’s job to one of task-giving and quality assurance (paired with targeted deeper work).
The average gross margin for a public software company is between 70-90%. That means they only spend 10-30% of revenue on actually delivering the product. The bulk of the remaining cost is Sales and Marketing — getting you to buy it!
If you strip the extraneous features and leverage AI for the build out, companies can make the in-house economics work!
What to replace
You don’t need to build everything in house. Please don’t.
The software you choose to build in house needs to be imperative to achieving a competitive edge in your market.
Typically these center around:
GTM / sales motions
Back office / operations efficiency
Client value delivery / customer service
Examples that I’ve seen recently and worked on:
GTM tooling leveraging proprietary data combined with public or paid data sources and internal secret sauce. Packaged together via AI and automations. Making lead generation and sales 10x more efficient than competitors.
Operations enablement software that eliminates 40%+ of workers’ back office tasks. Increasing efficiency and creating a hiring edge.
What changes are coming
More data hacks
Sadly, the rise of the “vibe coder” will lead to a lot more mistakes. Inevitably there will be more databases and APIs left exposed. The next 10 years will be a great time to be a hacker. This is a place where having a Fractional CTO or technical advisory is invaluable.
Earlier graduation from low-code tools
Make and Zapier have played an important role of bridging the gap between horizontal software and company-specific needs. These tools will still play a role, but we should expect companies to graduate to in-house applications sooner.
Developer tooling & infrastructure pricing models
Expect to see tools like Supabase deprecate or severely limit the free tier. Many companies could survive indefinitely on that tier today.
More “slop”
Slop means poorly designed software. There definitely will be more of this.
Controversial opinion: For internal tooling it doesn’t matter as much as people think it does.
Slop on non-critical surfaces can be ok. You definitely do want to have strong UI/UX support on core surfaces, though! 80% of the product can often afford to be “slop” as long as the 20% that provides value is amazing.
SaaS pricing pressure
SaaS tooling has always had to manage the “build vs. buy” equation to justify price. As the “build” side of the cost equation continues to drop, the price of buy will be forced to adjust as well. For any software you decide not to build in house, you may chose to use the threat of building in-house for pricing leverage.
Fractional developer boom
For many companies, it still won’t make sense to have an in-house development team. But, it will make sense to have a trusted agency or freelancer to build and maintain in-house software. I’m biased, but I strongly recommend working with a Fractional CTO to manage the build out including any other third parties.
What to do about it
If you own a company*, take a look at your SaaS spend. Compare it with the value you’re getting from AI tools like chatgpt. Think critically about what software or system could unlock more growth for your business.
Honestly, my guess is that you already have something in mind when reading this.
Build it.
Chat with your CTO (or loop me in).
Find a way to get it built.
Your competitor is building. You should be to.
* Company doing over $5m per year in sales (below that it may not yet make sense to build software in house).