Legacy software in seven years

Aakash Dhage

No-code automation comes back for another try every decade, but not because it didn’t work before. Like each new iPhone or new competitor, new business software regularly makes new attempts to solve old problems. No-code automation offers an especially good example, because advances is conventional programming over the last fifty years have done little to change how software development typically takes forever and costs all of your money.

In no-code automation, it doesn’t matter how many attempts it takes to make progress. We benefit from every new product that chips away that the mass of problems that still require conventional coding. For as long as we have too much code that needs writing, by too few programmers, who we struggle too hard to hire, market opportunities for alternatives remain.

No legacy

No-code automation requires compromises, and every product makes different trade-offs between expressive power (what you can build) and usability (how many people can successfully build). Even if you change nothing else, you’ll want to start again, and build a new product without the legacy, just to make alternative design decisions. And in automation, you can always try a different way to resolve the tension between tool complexity and audience breadth.

No history

Even if you think that all previous no-code automation products failed, or somehow have no value, someone with no history who’s never heard of them will try building a new one. Every year, we can find ourselves one university graduation project away from a new product category. Never underestimate what someone can achieve when they’ve never heard that it can’t be done. Especially when they have better timing.

New technology

Web applications have become progressively more graphical over the years, enabled by technologies like the HTML5 Canvas (2004), WebGL (2011), and WebGPU (2023). New technologies like these made a new generation of big canvas collaboration tools feasible. And each new technology brings us one prototype demo away another new product category.

New platforms

Less frequently, business software migrates to a new default platform, such as the 1990’s shift from Windows desktop applications to web-based software, and the later consumer software shift from web to mobile devices. Each new platform offers an opportunity to try the same product again, and find out how the new platform’s affordances and constraints change the solution.

If business software were to shift from web-based graphical user interfaces to AI-based text chat interfaces, spoken or typed, we might see a new kind of no-code automation. Out with the graphical process notation, and in with natural language domain vocabularies.

New markets

General purpose software, such as automation platforms, potentially adds value in any imaginable market. However, to make progress, start-ups initially focus on sales in familiar territory, geographical or otherwise: golf clubs in Portugal, or financial services in the UK, or manufacturing in Austria, for example. A product that didn’t succeed before might have more success with a different audience in a different market, before using that success to break out into larger markets, such as corporate America.

You’ve got seven years

I recently chatted with David Roberts, Head of Product at n8n, and he quipped that, ‘in automation you’ve got seven years to succeed before someone else comes along with a new approach’. I agree; a lot happens in seven years.