As I have been writing about adopting data-driven decision making in the broad sense and data-driven development more narrowly over the last weeks, several people have reached out to me, discussed with me and more generally explored the subject with me. During these engagements, I realized that many products and services exist for the sole purpose of creating a digital representation of a real world entity.
For example, a digital camera takes an image of a real-world object. A medical imaging product like an X-ray or MR scanner creates digital images of the insides of the human body. An electron microscope makes digital images of materials or biological tissue. A wearable collects heart rate, exercise and sleeping data, among others. Your mobile phone (and the infrastructure it talks to) knows where you are, who you talk to and for how long, which apps you use, etc. Your car knows where it has been, how long it has been driven, what the average speed was, the fuel usage, etc. Even a modern radar system is little more than a set of AD/DA converters that seeks to convert any signal into bits as early as possible.
All these products, and many more, capture and store the data in digital form. For the longest time, many of the people I meet as well as myself, thought of this data as an artifact in support of the physical product. Such as the car collecting data about itself allows for preventive maintenance or my wearable telling me to sleep more for health reasons.
The idea that I challenge you to explore in this article is to invert the above. Imagine that the sole purpose of all the products, systems, solutions and services that we use on a daily basis is to collect data. To act as sensors in the physical world in order to feed a virtual, digital universe. That the purpose of each car is to contribute to a virtual model of mobility patterns. For wearables to collect data from millions of people to establish correlations between exercise and health, between sleep and productivity. For mobile phones to collect consumer preferences, etc.
In a world where the amount of data is growing exponentially, both in terms of the amount of data collected in real-time as well as the storage of data, and where some futurists and venture capitalists predict that everything physical will cost $1 per pound, atoms are commodity and the bits are where the value is.
In several parts of our life, this is of course already a reality. Companies like Google, Facebook and LinkedIn give us access to their services for free because they monetize the data that they collect about us during use. Similarly, Amazon offers physical products like the Kindle or Fire basically at or below cost and with no or negative margin in order to monetize the digital books and other (digital) media content. Companies like Uber and AirBnB are valued so excessively high, by traditional standards, not because of the cars, drivers and rooms, but because they’re based on data and algorithms. In fact, an Uber driver or AirBnB host works for an algorithm.
Although I of course understand that far from all industries have made the transition from the analog to the digital or from the physical to the virtual, it is clear that companies that put data before the physical product have a number of significant advantages.
First, it allows for a continuous business model. When selling a physical product, you are by the nature of the business forced into a transactional business model where the customer buys the product, goes away and comes back years later to buy another instance of your product. When the business model is based on digital technology, it is much easier to establish a product-as-a-service model, usage-based pricing as well as other data-driven services as you know can compare the behaviors of thousands or millions of users.
Second, it allows for a multi-dimensional monetization approach. Once you are receiving data from hordes of users, this data can often be processed and made available to other parties. For instance, when collecting data on car usage, insurance companies may be interested in specific views on that data for improving models for setting premiums. When selling a physical product, you are typically stuck in a one-dimensional business model: you and your customer. This is a competitive risk, because the first digital competitor who figures out how to monetize the data through other ecosystem partners can start to subsidize the physical product and either push you out of business or force you to operate on razor-thin margins.
Third, it allows for continuous improvement of the product or service through continuous deployment of new software. When buying a physical product, we accept that it is at its peak state when taking it out of the box and that it will deteriorate over time. For digital products, the mechanical parts of the system may get worse over time, but the overall performance typically gets better over time. Periodically, I may be forced to replace the physical shell with a new one, but the overall experience is one where things get better continuously.
Fourth, once I have large amounts of data, I can use machine learning and deep learning solutions to make my product more intelligent, proactive and, in the end, perhaps even autonomous. Although many worry about the data collected about us on a continuous basis, the fact of the matter is that if we ever want products to act in a proactive fashion, to predict our next step and even to act on our behalf without us initiating the action, these products need to constantly monitor and observe us. Where else would the data required to proactively initiate actions come from?
Fifth, digital products and services allow for cross-product, cross-industry and cross-context connections that could never have been achieved through physical means. For instance, in a digitalized office building with access control, a security system, automated lighting and intelligent heating, the heating system can use the information about the presence of people in the building and specific rooms to allow for higher fluctuations of temperature in those locations where it knows, because of the access control and security systems, that nobody is present in specific rooms and offices. Visa versa, the security system can increase the sensitivity of its sensors and even use the sensors of other systems, like automated lighting, in order to improve its performance. And on top of all this, because of the flows of data, these systems can even learn, resulting in an office building that gets better every day that you use it!
Concluding, companies and industries exist along a spectrum ranging from 100% driven by atoms (physical world) to 100% driven by bits (data and virtual universe). Although the notion of data being the most important and thinking of physical systems as the means for creating a digital twin for everything and everyone in the world is an alien concept in many industries, the companies that make the leap into digital first gain enormous competitive advantages. Thinking about this is small steps is the wrong way to go about it as it will be way too slow. Instead, take the extreme position of the data and the digital twin of your customers, products, systems and solutions being the only thing that matters and everything physical being a means to that end. Based on that, determine the implications for your company and start executing on bringing that future about as rapidly as possible. Because if you don’t, I promise you that some new entrant in your industry will do so and disrupt you and all the other incumbents.