Jan Bosch is a research center director, professor, consultant and angel investor in startups. You can contact him at jan@janbosch.com.
Many companies are good at creating sustaining innovations in their mature, revenue-generating products. We know what KPIs customers care about and we focus our efforts on continuously improving those. For example, because of regulation as well as customer preference, the fuel efficiency of vehicles as well as their reliability have steadily risen over the years.
The challenge with digitalization is that the sustaining innovations that were successfully convincing customers to keep buying become less effective and customers want to see different forms of delivering value. As an example, the car I drive (a German brand) has lane keeping and adaptive cruise control and I use these features quite a bit. Unfortunately, the functionality isn’t very good and the car tends to act weirdly in specific situations. I’m OK with that. The thing that annoys me to no end, however, is that I know that the functionality will stay equally bad for the entire time that I will own the car. Unlike my computer, my phone and other equipment, it won’t get better!
Similar to how Nokia (for a wide variety of reasons that I won’t go into here) missed the transition in consumer preference from product variants to apps on top of a product with minimal variation, many companies risk missing a fundamental switch in customer preference due to digital technologies. As always, your customers won’t be able (nor is it their job) to tell you what they want, but they most certainly will recognize it when they see it and change their buying behavior. Of course, you can keep going on momentum, brand and customer relationships, but you need to adjust or risk disruption.
When it comes to digital technologies, there are at least three major technological approaches that need to be adopted: DevOps, A/B testing and artificial intelligence. The essence of digitalization is a fundamental shift in value delivery from transactional to continuous. For most companies, this can only be achieved cost-effectively by changing the software in the offering, rather than anything physical. Frequent updating of software in deployed products brings us to DevOps. For digitally born SaaS companies, this is obvious beyond belief and industry best practice for close to two decades now. For many cyber-physical systems companies, however, this is still a work in progress. There are many reasons, including regulations and certification as well as many of the company-internal justifications that we’ve discussed in this series of posts to date, but the fact that it’s hard is no reason to not get there.
Device lifecycle management for fleets of IoT devices
Microchip gives insight on device management, what exactly is it, how to implement it and how to roll over the device management during the roll out phase when the products are in the field. Read more. .
The second major technological approach is concerned with A/B testing and other experimental approaches. When we can deploy new software in systems in the field, we can also get data back from these systems. This opens up a quite significant shift in how we work with requirements and features as, rather than guessing about the value of new functionality to customers, we can actually measure it. By deploying small slices of new functionality in some systems and comparing the key KPIs between systems that have the new feature with those that don’t, we can quantitatively and statistically determine the impact. That allows us to stop the development of features that have no or even a negative impact and double down on the things that really move the needle in a positive way. For anyone who has been in feature prioritization meetings between product management and R&D, the idea that we can decide what to include based on experimentation instead of rhetorics and storytelling should come as a relief!
No post on technology-driven innovation can ignore artificial intelligence (AI) and this one is no exception. I’ve written about our work on AI and AI engineering in several earlier posts and my position hasn’t changed: machine and deep learning (ML/DL) offer fabulous opportunities for new forms of value. To function well, though, ML/DL requires data, and often lots of it, which requires the constant flow of data from systems in the field. Similarly, ML/DL models should be subject to the same DevOps cycle (often referred to as AIOps or MLOps) as all other software in our systems.
Most companies are very good at technology-driven innovation for their main revenue-driving products. With digitalization, however, the innovations that drove product success in the past need to be complemented or replaced with digital technologies and technological approaches. Three of the main ones include DevOps to continuously deliver value to customers, A/B testing to quantitatively validate the value of new features before building them and artificial intelligence as it allows for much smarter system behavior in a variety of contexts. As Tim O’Reilly said: what new technology does is create new opportunities to do a job that customers want done.