Digital transformation is very hot right now.
The tech world goes through fashions. Each fashion period lasts for a few years at most, and then the formally hot tech topic moves on to hopefully become the new normal. Of course, what drives the hype is the potential for profit margin by vendors and services providers, and this supports the funding for change agents to deliver ever more strident messaging on the belle de jour topic.
The best of these fashions become part of the way we do things evermore, and while the margins for implementers shrink, the technology becomes a steady value add in our lives. This cycle does not only apply to hard technology, but is equally true of broad practices. Service management was very hot 15 years ago. It may have lost its marketing edge, but virtually all tech uses now have some color of service management in their processes.
We are currently in the high season of digital transformation.
Automation, on the other hand, has been around for a long time and will continue to be a key tactic to deliver speed, scale or cost reductions to the diverse processes that deliver the built environment we live in today. But automation is also subject to modernization and improvement, and in taking evolutionary steps forward, organizations are being sold old school automation as digital transformation.
This article differentiates between automation, and digital transformation, and in so doing proposes where each adds the most value.
Automation without digital transformation
A key differentiator is required when discerning automation from digital transformation. This can be hard to do because most current automation is driven by software, and both automation and digitization strategies are delivered on computing platforms.
What could we consider pure automation? When I started my career in the early 1980s, I encountered an amazing machine. Its significant role was to orchestrate the purification of water to be added to the primary circuit of a nuclear power plant. As anyone who has dealt with water purification technology can attest, there are multiple phases in the operation of such a machine, with filters being backwashed, and valves being opened and closed at intervals, all to get the best water quality. But the most surprising aspect of this machine at the time was that it did not use SCADA or any other electronic controller, it used an electromechanical activated perforated sheet, just like those roll player pianos of the 1920s. An example of pure automation with no apologies to digitization.
What would a modern equivalent be? Of course, these are legion, but the newest and brightest examples must be the cluster of self-learning AI solutions that deliver logical, but morally unacceptable answers. Invariably, the owner of the problem is at a loss as to explain the logic that the AI applied to arrive at the awkward conclusion. The equivalence of AI bots to the electromechanical automation of a 100 years earlier is now obvious. We have a machine that does what it does, but we are unable to instrument it, trend it or otherwise manage it.
Many systems requiring human intervention during operations fit into this category. A system engineering process that needs human hands on management to deliver software changes is automation without digitization. A nightly backup where stats are monitored and entered by people is another. Unless there is time allowed in the process for updating metrics and performance indicators, the system becomes opaque to anyone or anything that wishes to see what is going on, irrespective of how much is delivered by automation.
Automation speeds up processes, but in doing so can limit the flexibility of the organization to adapt to change. Software driven automation is less rigid in this regard, but as seen in the AI use case above, is certainly not always malleable. Automation only supports digitization if it also delivers enough operational data to allow an abstract model of the system to allow scenario testing and dynamic changes.
Automation is ideal for systems where the value added is consistent, predictable and commoditized. Volumes should be expected to be significant, and cost of access to capital is a key differentiator to beat competitors.
Digitization without automation
What then might digital transformation be without automation? The most ubiquitous example is the humble mobility service. Uber, Lyft or Grab, pick your favorite.
The service is patently digitized. Your smartphone location is key to you calling your ride. The diver’s smart phone location is key to allocating the gig. Your payment details are already entered, and your payment for the ride is a wonder of digital payments. The cost of the ride is calculated not only by the time and distance to complete the journey, but also factors in the time of day, the relative demand for the service and multiple other external factors. Uber has clearly created an abstraction of what is going on and is clearly gaming the situations as they arise to deliver a good user experience and profit for itself.
But, and despite the clear moves to bring in autonomous cars, the ride itself is still manual. The car is still steered by a human driver. The driver still greets you, the driver still identifies you as the putative customer, you are provided with the vehicle details and the driver details. You still validate the diver as the service provider. It is still a human that does these basic security and safety checks. The actual mobility service is still not automated.
So here we have clear examples of a manual service that has been digitally transformed.
There are infinitely many additional examples.
The pizza shop that allows you to book when you want your over delivered or picked up. Here a very manual kitchen logs where each order is and prompts each station to do its bit to deliver performance data so that the order can be ready within its intended serving time slot.
The restaurant that manages your order, calculates you bill, while the waiting staff are free to interact with you instead of managing simultaneous servings for each table of guests is another example of digitization without automation.
Tracking your parcel from Amazon. While a multiple of different partners, automated or not move the goods about, they all log where your stuff is, and you, and Amazon watch and wait.
In all these manual digitized services, the processes are very carefully instrumented to provide users with timely operational data, and the business platform operator with rich business model defining information that enables detailed and accurate reappraisal of the optimal revenue possibilities for the business.
What makes a business digital is not the level of automation within its processes, but the capabilities of the business to adjust its key operations when the metrics delivered by the instrumentation within its processes indicate change is needed. While this speed of change is ideally real time, something of the order of days or weeks may be necessary to clear out existing processes. This business adaption is much faster than the speed of transformation attainable through business redesign.
The platform business as the current end goal of digital transformation
The above argument is that automation is not necessary for digital transformation. What is key is a well instrumented process delivering the metrics that support the right critical business choices. The ultimate business choice is the business model itself. In this world view, the value proposition of any activity is under constant review, and there may be several parallel activities delivering revenue in a digitalized organization.
At this point the platform business model becomes powerful. The goal of the business shifts to revenue rather than exploiting a particular product line automation or digitization. Each product line becomes its own abstraction, and each abstraction is constantly being updated by the metrics from its embedded instruments within its processes. Each line of business can be delivered internally or externally, partners can compete, and the focus can shift to customer experience and the attendant margin growth.
So, we can argue that to automate without enabling digitization is a strategic liability. Equally, we can argue that to consider digitization without aiming at an eventual platform model is to leave considerable opportunity unrealized and is also a strategic tragedy in the making. Of course, the goal is not necessarily to be the center of the universe platform like Amazon, but rather to build any value streams so that they can thrive and exploit the power of platforms to your additional benefit.
In every case then, the best goal of digitization must be to form a component in a platform business ecosystem. Conversely, the business case for automation is a more tactical one, and can be pursued at lower risk to an organization.