No-Code won't die, SaaS won't die

There’s a lot of talk about no-code and SaaS being redunant in the age of generative coding. I’m an unlikley defender of no-code, but they will live on, as will coding as a profession. As the tree is shaken it will become increasingly clearer what these different tools are for and where the limits or pitfalls lie.

First of all, “no-code” is a misnomer, it’s really “no-coder”. There is plenty of code. It’s not clear what it does and you don’t have access to it to check. That’s fine while it works, until it doesn’t.

Nothing to install. Your triple IT by-pass

This “No expensive coders or IT department between you and your dreams” sales pitch was pretty much how SaaS was sold into small teams. If a department head had a company credit card on which they could pay for 30 user accounts for a CRM or project management system they were off to the races. No legal, no IT, no custom code. If they were smart and bought into a startup SaaS you even help steer the product roadmap as early stage startups are in listening mode. As a buyer, I worked with several Customer Success Managers to lobby for features, even writing user stories so they could just slot them into their own processes more easily. CSMs seem to have gone by the wayside nowadays.

Then, as the last decade dawned, simple workflow automation apps (think IFTTT and Zapier) grew out of the need to connect proliferating SaaS apps. IFTTT came on to the scene in 2010, closely followed by Zapier in 2011. Salesforce kicked off the SaaS revolution in 1999, but it was the combined spread of data across disprate SaaS AND the availability of APIs for those SaaS that supported the growth of connectors and workflow apps.

No-code grew out of the need to move beyond these simple connectors. Founders and small teams couldn’t build what they wanted by cobbling together existing SaaS apps, nor should they, and boot/cashstrapped founders wouldn’t pay for custom development until they fleshed out their ideas and had some traction. So no-code offered a blank canvas on which to sketch out your startup’s app while maintaining the promise of SaaS; no IT invervention required. In fact, it was now being taken up by a generation who had no memory of on-prem installed enterprise software.

It’s all about shipping

This brings us to why people think that no-code will be dead by the end of the year and I don’t think it will.

No-code platforms are a trap for founders who do not see it as a step on the road to a robust online business. There’s an inflection point from initially feeling very empowered that tips over into the realisation that you are now building your business within the capabilities and limits of the no-code platform. That is coloured to various degrees by your understanding of what it can do, how things work on the web, data architectures generally and a willingness (or lack of) to pay for the more advanced features you’ll most likely need.

There’s also the founder trap that we saw in the 2010s where founders would go to coding bootcamps to learn to code and build their app themselves. This led people to be invested in what they had managed to build rather than advancing their actual business beyond at best mediocre design and a pile of technical debt. Founders who invest time in building their own solution find it difficult to kill their darlings. But, no-code has a massive advantage of actually being shipped and deployed from day one.

Vibe coding in a tool like Claude Code results in non-technical founders feeling like they can break out of those visual and technical design limits. But deploying is much harder for non-techs. Deploying safely and securely a rare thing.

If we’re honest, SaaS will do 80-90% of the job an enterprise needs and there’s a chance that staff have used the same system during previous employment.

No-code is OK for prototypes, a means of a non-technical founder walking through the destructuring their problem domain, of getting something in front of prospective customers and gathering a whole load of learning about your assumptions before commiting real money. It’s also perfectly fine for some internal company and some B2B applications you might even use it to fight your corner for some custom development.

So where does that leave us?

I thought generative coding in the hands of everyone would lead to an age of micro tools; tools built to solve the task in front of you and no more, the end of the mega-suites. That may still come, but not in this iteration. Conversational generative coding is now available in No-code, it’s also in SaaS. That will keep them alive and improve them.

But I still believe that ultimately, voice will be the interface, not your grids and forms on the screen, not your PPTs and PDFs.

In the mean time, we’ll continue adjusting our roles, over estimating the degree to which our expertise matters to other people.

Founders will continue to become enamoured with what they’ve managed to build at the cost of focussing on customers.

Developers over estimate their ability to get buy-in for the big picture; the importance of deriving value from data later on in its life and preventing data fragmentation.

Users will remain task achievement focussed even though they understand that they really shouldn’t export that attendee list to excel to fiddle and filter it for their next event, but …

Designers will bring value, but because great design is something you don’t notice most clients will see the job as done 80% of the way along the road.

UX people will bring massive value to a project but will probably under estimate the amount of pain and frustration a user will tolerate to get something done.

And so we go on reinventing in an infinite loop.