There is a ideal story: If Low-Code development tools can be 10x more productive, then we gain competitive advantage over the cost of any kind of software project. Outsource vendor will want to purchase such kind of tool to compete with others.
Why this ideology does not work?
- Unlike SaaS model, user will find the vendor through market search. Outsource vendor seek customers actively, it does not have the luxury to filter out "good quality" customer through funnel. To keep the utilization rate, sometimes some less ideal project must be taken. Those newly found customer will have high business development cost. Also some revenue will have high cost to collect after the project finished a long time. Among all the cost of outsource project, high development labor cost is just one, sometimes the insignificant one.
- Tool can not be fully automated. We should need human to drive the tool. Human need certain period of learning, and need to make design decisions wisely in daily use. Even if the tool is bug-free, and well-documented, learning still takes time. Also to archive engineering property such as "low coupling, high cohesion" is still a black art largely depending on the human designer, no matter what the development tools market itself.
- Development tools users are developers, SaaS users are normal corporate worker. Developer will be more concerned with vendor lock-in. Why? To SaaS user, it is just a matter of choice, they will be lock-in by you or others. But to developer, they will think open source developer tools is truly in their hands.
- Developer will have career concerns. However, this is a double edged sword. Like salesforce admin will try their best to maintain salesforce in the market, given their investment in the technology.
- I am not saying open source === no profitable business opportunity. There are many successful commercial company behind open source projects. But in the area of development tools, it is more likely a consulting model business. It might be profitable, but due to human intensive nature, it will be unscalable. Investing millions of dollars on Low-Code to sell copies of the tool will likely lead to commercial failure. It has a much better chance of success running as not for profit open source project, like Ruby on Rails. Next DHH could be you.
- The tool will be easier to sell to have immediate effect on existing codebase. I have written a record & replay tool that could re-produce production error without code modification. The tool took the heart of developers over-night. Low-Code does not work for legacy code, it targets green field projects. It is much much harder to sell if the tool can not solve any existing problem just now. It takes courage to trust something might work in the far future.
- 10x more productive does not mean 10x drop in latency (time to market). Unlike outsource vendors, SaaS companies, Internet business does not care the total cost that much. However, extreme time to market is desired, in hope of market domination. Modifying PHP file in production is cutting corner, but maybe exactly what is needed to be extremely fast.
- Is 10x more productive Low-Code a lie? I still believe it. But it will not a good business commercially given the points above.
There are many startups in this area recently. If not selling development tools directly to developers, what market should they target?
- SaaS + lowcode: Besides "truly useful to the user", stability, speed of page loading, the 4th should be "on premise deployment and customization". SaaS can sell to more customer if have Low-Code capability built-in.
- lowcode + a family of SaaS: Instead of selling as development tools, lowcode can be wrapped as a family of SaaS. It will not be that feature rich (or feature bloated should we say) to be competitive by itself. But it will be appealing to those who want customization.
- Data Centric base: Microsoft launched a product Dataverse recently. It might address the problem described by the book "Software Wasteland". I would describe the solution as "Backend as a database" for the age of frontend-backend separation.
You can also find me at twitter https://twitter.com/nctaowen