Why telecoms business transformation is hard

Telecoms is one of the last industries to move from a “supply push” model to a “demand pull” one. We’re lagging because the task is particularly hard.

I am relaxing in the plush home of family friends in Edinburgh. This property was bought from the proceeds of a highly successful venture in compostable catering supplies, where one friend is CEO and founder of Vegware Ltd. The differences between his business and the start-up where I am co-founder are instructive.

With Vegware, the essence of the business is import and distribution. New materials (“biodegradable plastic-free plastics”) had already been invented. The machinery for making disposable forks and cups has been around for decades. Container shipping from Chinese factories to European markets has long ago been perfected. The goods are bought and used in the same way as before by customers. The management theory being applied is all well established.

The result, however, is transformative to the sustainability of the catering business. Rather than having waste that will sit in landfills for thousands of years, poisoning the environment for generations to come, it becomes a closed system. The waste can be collected, composed, and turned back into the raw material for growing food.

In contrast, transforming telecoms requires a large number of novel and innovative activities to line up.

We’ve not yet applied the concepts of ‘lean’, as we have to re-interpret them all for distributed computing. In a Toyota factory, you can press a red “stop” button to halt production if you see a quality defect. In telecoms, the “cars” have already left the “information factory”; you can’t move quicker than the think you’re producing.

The concepts of physical supply chain management have not been consistently and knowingly applied to telecoms. Our resource scheduling and planning is very primitive when compared to the sophisticated logistics industry.

In order to take control over the supply chain, we need to deal with the fact that the customer experience is being continuously manufactured. Rather than discrete objects being shipped, we need to maintain an instantaneous service quality over time. This means we need new technologies and mechanisms for measurement, modelling and management of instantaneous quality.

Those in turn require a new set of network performance engineering methodologies to quantify supply and demand, and the “safety margin” between them, for any given customer demand and use context. These have to be adapted from other disciplines, such as safety-critical systems and supercomputing.

For that experience engineering to be possible, there has to be an underlying science, which is also new and not yet in the textbooks. It requires us to have a general theory of networks as “information photocopiers”, and how the network performance is related to the user experience.

Unpacking this, we need yet another conceptual change. In order to be able to control quality, we need to switch from a bandwidth-centric resource model (which is always about averages), to a latency-centric one (about instants). So the very foundations of the discipline need attention.

For that to work, we have to address an even more elementary issue. Because packet networks are “stochastic” (i.e. evolve over time through random processes), we need a new branch of mathematics to cope with the fact they have both packet loss (a discrete outcome) and delay (a continuous one).

So the “treasure” is ∆Q, and it is very well buried!

As an informed friend reminded me last week, in user interface design there is a general principle that ordinary humans struggle to cope with more than two layers of hierarchy. In the hunt for the missing telco treasure, we have at least six. Therefore it is unsurprising if most explorers give up looking for the prize long before they get close to it.

If Vegware had to invent the organic chemistry, develop the bioplastics, conceive of container shipping, as well as join all of these together, then it would have been an entirely different kind of enterprise. Yet that is essentially what we face in telecoms, plus the cloud application access goods are sold in a new way, too.

Our industry is presently mathematically unsustainable, and this shows up in its poor customer satisfaction ratings and weakening returns on assets. It has to face up to the depth of its problems, for progress isn’t possible without fixing the fundamentals.

The prize for whoever makes it first to the treasure and brings it to the surface is essentially to own the whole industry. This explains why I continue to excavate furiously, albeit with a shovel, when what the industry needs is a scientific opencast mine.