We are often talking about “innovation” and “digital innovation” (or “technical innovation”) in particular, when it comes to tech startups. It has, unfortunately, become a cliche, and now “innovation” is devoid of meaning. I’ve been trying to put some meaningful analysis of the “innovation landscape” and to classify what is being called “innovation”.

And the broad classification I got to is “technical innovation” vs “process innovation”. In the majority of cases, tech startups are actually process innovations. They get existing technology and try to optimize some real world process with it. These processes include “communicating with friends online”, “getting in touch with business contacts online”, “getting a taxi online”, “getting a date online”, “ordering food online”, “uploading photos online”, and so on. There is no inherent technical innovation in any of these – they either introduce new (and better) processes, or they optimize existing ones.

And don’t get me wrong – these are all very useful things. In fact, this is what “digital transformation” means – doing things electronically that were previously done in an analogue way, or doing things that were previously not possible in the analogue world. And the better you imagine or reimagine the process, the more effective your company will be.

In many cases these digital transformation tools have to deal with real-world complexities – legislation, entrenched behaviour, edge cases. E.g. you can easily write food delivery software. You get the order, you notify the store, you optimize the delivery people’s routes to collect and deliver as much food as possible, and you’re good to go. And then you “hit” the real world, where there are traffic jams, temporarily closed streets, restricted parking, unreponsive restaurants, unresponsive customers, keeping the online menu and what’s in stock in sync, worsened weather conditions, messed up orders, part-time job regulations that differ by country, and so on. And you have to navigate that maze in order to deliver a digitally transformed food delivery service.

There is nothing technically complex about that – any kid with some PHP and JS knowledge can write the software by finding answers to the programming hurdles on Stackoverflow. In that sense, it is no so technically innovative. The hard part is the processes and the real-world complexities. And of course, turning that into a profitable business.

In the long run, these non-technical innovations end up producing technical innovation. Facebook had nothing interesting on the technical side in the beginning. Then it had millions of users and had to scale, and then it became interesting – how to process so much data, how to scale to multiple parts of the world, how to optimize the storage of so many photos, and so on. Facebook gave us Cassandra, Twitter gave us Snowflake, LinkedIn gave us Kafka. There are many more examples, and it’s great that these companies open source some of their internally developed technologies. But these are side-effects of the scale, not an inherent technical innovation that lead to the scale in the first place.

And then there’s the technical innovation companies. I think it’s a much more rare phenomenon and the prime example is Google – the company started as a consequence of a research paper. Roughly speaking, the paper outlined a technical innovation in search that made all other approaches to search obsolete. We can say that Bitcoin was such an innovation, as well. In some cases it’s not the founders that develop the original research, but they derive their product from existing computer science research. They combine multiple papers, adapt them to the needs of the real world (because, as we know, research papers often rely on a “spherical horse in vacuum”) and build something useful.

As a personal side-note here, some of my (side) projects were purely process innovations – I once made an online bus ticket reservation service (before such a thing existed in my country), then I made a social network aggregator (that was arguably better than existing ones at the time). And they were much less interesting than my more technically innovative projects, like Computoser (which has some original research) or LogSentinel (which combines several research papers into a product).

A subset of the technical innovation is the so called “deep tech” – projects that are supposed to enable future innovation. This can be simplified as “applied research”. Computer vision, AI, biomedical. This is where you need a lot of R&D, not simply “pouring” code for a few months.

Just as “process innovation” companies eventually lead to technical innovation, technical innovation companies eventually (or immediately) lead to process improvements. Google practically changed the way we consume information, so it’s impact on the processes is rather high. And to me, that’s the goal of each company – to get to change behaviour. It’s much more interesting to do that using never-before-done technical feats, but if you can do it without the technical bits (i.e. by simply building a website/app using current web/mobile frameworks), good for you.

If you become a successful company, you’ll necessarily have both types of innovation, regardless of how you started. And in order to have a successful company, you have to improve processes and change behaviour. You have to do digital transformation. In the long run, it doesn’t make that much of a difference which was first – the technology or the process innovation. Although from business and investment perspective, it’s easier for competitors to copy the processes and harder to copy internal R&D.

Whether we should call process innovation “technical innovation” – I don’t think so, but that ship has already sailed – anything that uses technology is now “technical innovation”, even if it’s a WordPress site. But for technical people it’s much more challenging and rewarding to be based on actual technical innovation. We just have to remember that we have to solve real-world problems, improve or introduce processes and change behaviour.