Every company I meet claims that they’re agile. When I push a little bit and ask for details, typically the conclusion is that this is only true at the team level. The rest of the organization is stuck in traditional, slow cycles. The defense that I get is that the company is fast enough in responding to changing customer needs as they would have gone out of business otherwise.
To me, this is a typical case of “what got us here won’t get us there.” When reflecting on how we deliver value to customers, we can see a continuous acceleration of value delivery cycles. Initially, most companies deliver value through product generations. Every few years, the next version of the product comes out and people update from the previous version to the new one. Especially in software, but also for other technologies, the yearly update has become the norm in several industries, especially in IT. Now, we’re moving towards DevOps, DataOps and MLOps and we deliver value every 2-4 weeks.
Of course, the speed-up of value delivery won’t stop there. Especially in the online world, techniques such as A/B testing are used to even more quickly deliver value to customers. And, as a next step, some companies (and researchers such as in my group) are experimenting with techniques like reinforcement learning where systems fully autonomously experiment with their own behavior to optimize their performance.
To survive and thrive in a world where customers expect continuous delivery of new value, it’s necessary that software R&D teams are agile, but that’s far from sufficient. In this world, every function, department and team needs to operate based on Agile principles. In practice, the functions that struggle the most with the transition to a continuous value delivery are sales, customer support and mechanics and electronics R&D. We’ve already covered sales and customer support in earlier posts in this series.
Traditionally, the R&D teams concerned with mechanics and electronics often have a SOP (start of production) mindset. This means that they’ll be focused on working towards a specific date, typically rather far out into the future. At this date, everything needs to be in place, tested and guaranteed to work as it should as changes post the SOP point tend to be prohibitively expensive.
The interesting observation is that there’s a fundamental shift between traditional and digital companies: for the former, everything stops at SOP whereas for the latter, everything starts there. Once we start production, we can get products in the hands of customers and only then, the data starts to flow that allows us to start the continuous value delivery.
Several of the companies I work with have changed the R&D process from a product and SOP focus to a stream model where each discipline develops a constant flow of improvements to the subsystems or components that it provides to the product portfolio. Any specific product is then a composition of the selected variant for each component and subsystem and almost a ‘by product’ of the stream-driven R&D process.
This approach supports the notion of continuous value delivery as the new versions of components and subsystems may also be available to systems already out in the field at customers. So, we can deliver additional value through new and improved mechanical and electronic components as well.
The main precondition for this to work, though, is a strong system architecture discipline in the company that embraces the digitalization principles and enforces backward compatibility not just for software but also for mechanical and electronics components. And, of course, it calls for a well-managed evolution of interfaces where the old ones are no longer fit for purpose.
Most companies claim they’re agile, but in practice, that’s only true for the software R&D teams. The other functions and departments are still using traditional, slow processes. Digitalization requires everyone in the company to adopt Agile principles. The incarnation of Agile may look different, but the overall goal is to enable the continuous delivery of new value to customers through all means possible. Digitalization requires Agile for real!