SaaS is dead. Or at least, the model that’s dominated tech for two decades is on life support, and the plug is about to be pulled.
I came to this conclusion looking at my own P&L. My agency, Creative Marketing Group, had a software bill topping £30,000 a year. A graveyard of SaaS subscriptions. Analytics platforms, reporting dashboards, SEO tools, outreach automation, social media schedulers. Each one bought with the promise of solving a problem, each one now just another login, another dashboard to check, another workflow to manage.
We were sold a lie. The lie of SaaS was that by giving us tools, it would make us more productive. The reality is that it just made us busier. It gave us fragments of a solution and left the hardest part—the integration, the interpretation, and most critically, the action—squarely on our shoulders. It sold us a toolbox but expected us to be the master craftsman.
The next wave of technology fixes this. It’s not another tool. It’s an agent. It doesn’t give you a dashboard; it delivers the outcome. And if you’re still selling dashboards in the coming years, you are going to go out of business.
The fundamental flaw in the Software-as-a-Service model is fragmentation. It takes a single desired outcome—say, "increase qualified inbound leads by 20%"—and breaks the required workflow into a dozen expensive, disconnected steps, each managed by a different tool.
Your SEO tool identifies a keyword opportunity. Your content optimisation tool helps you write an article. A separate analytics platform tracks its performance. Another tool monitors backlinks. A social media scheduler pushes it to LinkedIn. A CRM, hopefully, captures any resulting leads. This isn’t a seamless workflow; it’s a digital gauntlet. The user becomes the human API, patching together disparate systems and manually executing the next step.
Consider the sheer logistical nightmare inside a major UK enterprise. Imagine the martech stack at a company like Tesco. They are likely managing hundreds of software tools. The average enterprise, according to recent data, uses over 110 different SaaS applications. That’s 110 logins, 110 billing cycles, and 110 sources of data that rarely speak the same language. The result is not efficiency; it’s chaos. It’s expensive teams of smart people spending their days exporting CSVs and wrestling with integrations instead of doing the high-value strategic work they were hired for.
The "Modern Data Stack" Is an Admission of Failure
Here’s my contrarian take: the entire concept of the "modern data stack"—all those Snowflakes, Fivetrans, and dbt models—is not the elegant solution its proponents claim. It is an enormous, expensive, and complex admission that the underlying SaaS model is broken. It’s a sticking plaster, a multi-million-pound workaround for the fact that the tools we pay for cannot communicate or act on their own intelligence.
We’re spending fortunes building data warehouses to unify data from tools that should have been integrated from the start. We’re hiring data engineers just to clean up the mess created by a thousand different apps outputting a thousand different data formats. This is not progress. It’s piling complexity on top of complexity. This huge ecosystem only exists because SaaS failed to deliver on its primary promise: making work easier. This realisation is at the heart of the collection of my thinking I