The End of Marginal Tech

Using software to conduct a business doesn’t make that a tech business.

It was 1984, when my mother, who was a manager at a bank at the time came home and told us that she’d be away for two weeks training for their new IT system; probably a massive IBM mainframe deployment, model S70 of the 80s or the like. Their current system was 100% analog with one computer doing menial operations, so it meant the end of mnemonics to remember the account numbers of the biggest and most profitable clients or spending hours on the phone to validate checking deposits. It was a lifetime ago yet still vivid in my mind, as I was the little messenger who carried the paperwork between the cashiers for pocket money in the afternoons in her branch.

Fast forward to the 2020s, every decent financial system from central banks to POS machines are connected to each other, nothing exists outside the realm of the kingdom of software. It’s not just finance obviously, everything’s online, every business runs on software unless you’re a solo gardener or a hairdresser, an artisan of sorts.

Let me indulge you with another personal story. A few years ago, we founded a startup which utilised creators (influencers) to do direct market research with their audiences, so instead of promoting products we worked with them to collect consumer insights. Our beautifully designed and developed proprietary software connected to the creators’ channels on various social platforms and we used the state of the art pre-trained AI models mainly by Open AI to analyse the conversation data. Our endeavour didn’t take off despite our utmost efforts mainly because the whole thing depended on relationships, human to human communications. We quickly realised that no matter how automated the system is, as long as you have creators in the equation you need humans to manage that business; they needed to be personally contacted, a trusted professional relationship had to be developed and maintained both with them and their managers.

In the end, our so-called tech business turned out to be yet another agency in the marketing universe and we decided to shut it down. We were a startup, raised investment, did great business with big brands but we weren’t a tech company. What were we then? Data; we were a business that collected proprietary consumer data for brands as a B2B SaaS (software as a service) company.

Our experiment isn’t unique, 47% of venture capital funding in 2023 went to companies with a SaaS business model and the global SaaS industry is worth more than US$237 billion.

Are these other SaaS companies tech businesses I hear you asking? Imagine an online booking system to manage reservations for a restaurant or a gym club, using a mobile app these companies make it easier for their clients to make bookings on their own without the hassle of talking to an agent. The company who provides these systems to the restaurants and gyms is now a SaaS business and is considered a tech company, yet they are software enabled booking managers. The actual value that comes out is data, all these companies do is collect massive amounts of data. In a world where frontier AI companies fast run out of available data to train their models, these so-called tech companies matter as they by the nature of their business collect and sit on precious, proprietary data.

The users who make these bookings for a yoga class or a dinner for a Friday date night are the product as usual; they not only pay for the service they’re getting but share their usage data for free with the vendor on top of it. Here’s a thought experiment, on its own the fact that I go to gym 5 or 6 times a week may not mean anything more than a simple data point to optimise the usage at peak times or maybe sell merchandise for a gym club, but when it’s merged with the nutrition data that comes, say from an instant food delivery platform tells a ton more about who I am as a consumer. Imagine both businesses are acquired by the same PE (private equity, which happens a lot from FMCG to insurance nowadays), and by the grace of six figure salaried lawyers every privacy law becomes malleable to manipulation and you got the holy grail of consumer behaviour data in your hands.

Or, the same PE invests in a co-working space and starts collecting data on who works for whom, who’s hiring, who’s firing, who’s expanding the office space, who’s renting meeting rooms with external guests who work for brands and so on and you’re ready to turn the whole business into an investment platform for angels with a recruitment or a credit business on the side. Hook the tenants on numerous subscriptions to ensure a steady flow of data and sustained month over month revenue stream and you’re golden. But this is not a tech business either.

In 2023, the average number of SaaS apps a business uses reached 371 and growing. Companies hire people to manage their subscriptions while having zero control over how these systems operate based on which algorithm as everything including their precious data sits in the cloud.

The whole show is about how to get data and marginal tech enablement is the vehicle to reach that goal. That’s what automation is about and AI will be the never satiated engine to eat it up.

What’s the real tech business then? What should the smartest kids work on instead of yet another marketing tool or solve non-existent problems? How about building life changing technologies; how to clean up our oceans, develop vaccines for cancer, invent better and cheaper solutions for offshore wind turbines… use AI to make them smarter without turning it into a goal in itself. This is a lot harder to do and requires a steel stamina to persevere in doing what you believe in and pitch to the investors who have the appetite of a monkey vying for the highest ROI in the shortest time frame. Yet, still, do it, it’s much better for growing your tech talent and helping humanity at the same time.

This content is not assisted by AI.