The Case for Custom Software
If you really want to automate your business, you'll only get so far with SaaS
See ya later SalesForce, Quickbooks, Zapier, Netsuite, Intercom, and every other major horizontal SaaS company, you had it good for a while.
SaaS thrived on the promise of scalability – distributing identical code to millions of businesses through the magic of the internet. But this model has a critical flaw: it can't perfectly solve any single business's unique problems. As code becomes cheaper to produce, the balance is tipping in favor of tailored solutions.
The power once held by SaaS giants stemmed from the mountains of code they'd written—a moat that's now eroding. Whether AI will churn out high-quality code next year or in a decade is debatable, but the trajectory is clear: we're sliding down the slope of ever-decreasing software costs.
This seismic shift opens up a world of possibilities for businesses of all sizes. Custom software, once the domain of tech giants and Fortune 500 companies, is becoming accessible to a broader range of organizations. Here's why this matters and how it's reshaping the business software landscape:
The Inherent Limitations of SaaS
When you target a large addressable market, you're not really solving any business problems perfectly for any of your customers. Sure, there might be some configurations you can do, but there's no place in the QuickBooks configuration to tell it, for example, that Tara said invoices must not exceed $50k per month, so Lynette should track any overage in a separate spreadsheet and add it back in some month when the invoice is less than $50k. Businesses are full of little rules like that. The bigger the business, the more there are, and SaaS sucks at dealing with them.
This fundamental limitation of SaaS is becoming increasingly apparent as businesses grow more complex and demand more tailored solutions. Forward-thinking companies are starting to look elsewhere.
SaaS is eating itself from the inside out
Meanwhile, it seems like SaaS companies know this. They’re sprinkling AI on their products like a 3 year old with a cake and jar full of jimmies. So far, these AI features don’t provide much business value. Since their connection to other systems is rigid and knowledge of data in other systems is tiny, there’s not much they can offer in terms of automation without having to write some custom code.
But if they don’t solve every problem, at least we can trust they’ll be here for the long haul? Not so fast. We didn’t always trust in SaaS. In the mid-2000s, IT managers had to be convinced that it was safe and secure and here to stay. Now, though, even Google is fighting for its life against OpenAI, Perplexity, and itself (Gemini). In addition to big threats from AI, there is never ending, market eroding, competition in SaaS. Vertical SaaS for tattoo shops, gyms, small grocery stores, large grocery stores, golf courses, everything are eating away at the addressable market for horizontal ERPs. There will be further and further verticalization of the various SaaS product markets in the coming years. When it becomes possible to have a viable business on a customer base of 10-15 companies, people will do it—they already do it!
Custom software is better for people
Businesses are a way of aligning groups of people from employees, to customers, to investors around a shared goal, and software is a critical component for scaling those organizations. Modern software has enabled organizations to effectively grow to millions of people. Walmart has over 2 million employees! But everyone who has ever worked for one of those organizations knows that huge companies are heinously inefficient. Forgetting to put the new cover sheets on your TPS reports can be soul crushing.
Small organizations that let humans work as people doing people things are way more efficient at making decisions, changing their minds, and getting work done. Part of the problem with big organizations is that even though they make good use of software to keep everyone marching to the same beat, the software forces people to behave in certain ways and it becomes the job of many people within the organization to make sure that people are conforming to the communications and process policies and constraints created by the sheer size of the organization and the software it uses.
This is where custom software shines. While it may require more upfront investment in time and resources, it allows organizations to build systems that work with their unique processes, not against them. Instead of forcing people to adapt to rigid SaaS solutions, custom software can be designed to enhance existing workflows, potentially reducing long-term costs associated with workarounds, integrations, and unused features.
If a 100,000-person organization could somehow be a highly orchestrated hive of 5000 twenty-person organizations, with each of those smaller organizations using the effective processes and human communication afforded at such small scales, we could do away with huge swaths of management and mind numbing software. Custom software could be the key to achieving this level of flexibility and efficiency, adapting as quickly as the business itself evolves.
Embracing the Custom Software Revolution
I think this is the promise of custom software. I can imagine a world where spreadsheets are ok, but automated, where the integrations and inputs and outputs across teams can be put into code almost as easily as you can think it. But in the meantime, it’s time to start planning for that future. At the very least, I suggest if your team is evaluating RAG solutions or trying to figure out how to ‘use this AI stuff’ to do something useful that you don’t go looking for a SaaS product to help you do it. Build it yourselves. Own it, and improve it. Take a ride down the slide of the ever-decreasing costs of software and land in the future.
—Jon Christensen
ps. Kelsus would happily do a free evaluation of your business or organizational unit to see where custom software makes sense. Reply to this email, and I’ll set something up.