Continuous dialog

During the Christmas break, I read several books including “Sense & respond” by Jeff Gothelf and Josh Seiden. The authors frame the difference between traditional and digital companies in terms of the absence or presence, respectively, of a continuous dialog with the market and the customers. Of course, traditional companies interact with the market and talk to customers, but once a product specification has been defined, the product developed and the product is ready for sales, the discussion is mostly concerned with getting customers to actually buy the darn thing.

If the sales are disappointing or if there are many requests for changes, there often is an upgrade or next version of the product. However, the frequency of updates tends to be low and is often measured in years. In automotive, a car model usually lives for 5-7 years and has a mid-life upgrade after 3 years. Even traditional software products experience a yearly release cycle at most.

Digital companies, on the other hand, are in a constant dialog or conversation with the market. Through a variety of different metrics and mechanisms, they continuously learn from the feedback from customers, make changes and measure their effectiveness. These mechanisms include instrumenting the product to understand how the product is used, the ability to provide different feature versions and ways to assign users to different groups. And, of course, digital companies assume continuous deployment (DevOps) to be in place as the feedback cycle should be as short as possible and preferably be measured in days and weeks.

Another main difference between traditional and digital companies is that the former tend to focus on what customers say while the latter focus on what customers actually do. This is a very important distinction as, although we all view ourselves as rational people, it turns out that in almost all cases, there’s a significant gap between what people say and what they do. Deciding based on what people say can therefore really throw you off the optimal path as the things you build or add to your product may not deliver any actual value.

Going from focusing on what people say to what people do can be a difficult transition, especially in B2B contexts. Customers may walk as they feel you’re not listening to them. Also, it’s really hard to tell them that they may believe something about their behavior that’s actually incorrect. However, the alternative is to build functionality that won’t add any business value in the long term.

Especially for embedded-systems companies, there often is a belief that these principles only apply to software, but not to electronics and mechanics. The interesting development is that more and more companies are now starting to explore mechanisms to deploy new electronics and, perhaps, even mechanical components to products in the field. Of course, the release frequency is lower for anything that includes atoms, but it’s entirely possible to also have a dialog with the market for these types of components. It’s mostly a matter of finding the right business model to support it.

Many in R&D don’t realize the economics of R&D budgets. For a company that allocates 5 percent of its revenue to research and development, it means that every euro invested has to generate 20 euros of business value. It’s not enough that the R&D costs are covered. The best way to ensure that R&D activities deliver the 20x return is to take small steps, measure continuously whether you deliver on the expectations and change immediately if the response from the market is different than expected.

Digital companies are in a continuous dialog with the market, have short feedback cycles and focus on customer behavior (rather than on what customers say). As digitalization continues to be the theme for 2021 as well, evaluate your business and technology strategy to ensure that you’re incorporating these principles and transitioning towards these. The only constant is change and not changing has a very predictable outcome: extinction. And who wants that?