Notes from Digital Transformation World

I am on my way back from attending TMForum Digital Transformation World Ignite! and I thought I’d share a few thoughts I had about what I heard while I was there.

Firstly - I’ve been working in telecommunications for almost 20 years and this is the first time I’ve attended DTW. I was certainly struck by how many talented and caring people we have in our industry, it was a great place to be for a few days. I certainly hope to be able to do it again.

Some background

I entered the workforce in 2005, and my introduction to our industry was the swan song of the 2G era, the start of the transition from voice and SMS to packet networks and data being the main service users were consuming. This is the time we were facing into the serious challenges of our commercial models - significant increases in the volumes of data we had to carry, and a linear increase in cost as we scaled up. This was at the same time as our customers were expecting flat or reducing prices, and we were fighting with our competitors, which meant it was something of a race to the bottom.

As I reflect, I consider that in many ways the challenges we face now have not changed since those days in 2005:

  • At the network level, our costs are still somewhat proportional to the amount of traffic we carry, and these costs are increasing as we carry more traffic
  • Our revenue is declining in real terms
  • We have a huge proliferation of complexity, some of which is structural (as our networks become more complex) and some is self-inflicted (through poor integration after M&A activity and under investment)

The chart below, taken from a recent OECD report into financing broadband networks shows how in the period from 2008 and 2022, European Telco stocks declined by 47% and those in the US grew by 33%, this is compared to the Dow Jones Industrial Average growing by 159% in the same period. This is not a perfect representation of our issues, but if we accept that stocks 'price in' the economic situation a company faces, it is a reasonable proxy for our issues. The OECD report is here.

View of telco stock performance

So, if our problems today are the same as those of 20 years ago, why have we not solved them? I have a few theories on this:

  1. Telcos had the right idea, attempting to move into adjacent spaces, however the rise of the digital natives was not foreseen, and these companies have captured a lot of the value which is dependent on the (somewhat more mundane) underlying network
  2. We have some embedded ‘slowness’ from our need to globally integrate. Telcos cannot all act independently, it would be chaos if we all did things differently (think railway lines with different gauges). Contrast this with companies like Netflix or Spotify, if they want to make a change to their protocol, they just need to negotiate with themselves!
  3. Many of our moves to capture value have been centered around financial engineering, for example the InfraCo, NetCo, ServeCo splits - and while these might generate temporary shareholder returns, they tend to embed further complexity into our business operations

Given this context, here are my thoughts on the ‘missions’ from TMForum.

DTW missions

Composable IT & Ecosystems

This is something I’ve been a proponent of for many years, not just for Telco, but for all organisations which cannot exactly predict how their technology will need to be used in the future.

I consider this is critical, but it is not easy, and my view of the critical success factors are:

  • We need to take a business capability led approach to allow us to deploy fewer, simpler systems. This allows us to find where we have duplications in our IT estate which can be eliminated, leading to less ambiguity and faster development
  • Standards based interfaces and data models need to be embraced. This is a critical role bodies like TMForum play, with tools like ODA and SID, which give us optionality to swap out/mix-and-match between systems and applications as we need
  • Having a north-star. This might sound like a no-brainer, but I’m constantly surprised by the amount of ‘use-case’ driven work there is, where no specific target or north-star is driving how we get to a specific systems outcome. We have to keep this in mind, and make an effort to get there.
  • Embracing self-service. There have been significant improvements made in self-service capabilities recently, much of which has been enabled by Generative AI. Is is a very useful tool which will allow non-technical people to build sophisticated pieces of technology. In the past, such things have been seen as ‘shadow IT’. This is something to be avoided, however there is room to enable business experts to build their own tools within specific guardrails.

Autonomous Network Operations

This is an area of real promise, however I think there’s a disconnect between the technology and how it will translate into real cost savings.

As an industry, we spend most of our money on capacity expansions, operating the network and dealing with our customers. The proportions of these cost categories vary depending on our role (ServeCo, NetCo or InfraCo) but they are typically the main buckets for all of us.

My view is that there is little fundamentally new technology being considered in this domain, and some concepts are already embedded in common network protocols - consider BGP (Border Gateway Protocol), where each network is referred to as an ‘autonomous system’ (AS). Despite the fact the technology is not new, it has not been widely adopted and we need to look into why this is the case.

Some barriers we’ve encountered include:

  • Human factors: people can become set in their ways and find it hard to understand how they can adopt new technologies.
  • Process factors: real cost reduction tends to come from changing processes, but often times we neglect the link between the technologies we deliver and the processes they are intended to change.
  • Data factors: this will come up again later, however, automated decisions are only as good as the data we put in, and when that data is poor it can lead to poor decisions, and this can undermine confidence in technology.
  • Cost factors: traditionally there has been significant cost to building rule based automations, and these have turned out to be brittle. Companies have found it easier to deal with the same challenges by scaling their offshore workforces and this has proved more cost effective than building automations

We have to tackle these factors equally along with the technology in order to make a real difference.

Data & AI Innovation

This is the fundamental enabler which powers much of the change we need, however I often notice the focus is in the wrong place. Telcos and our vendor partners tend to focus on the use-cases and the new pieces of technology and then neglect the base currency it all runs on: data.

Automations, no matter how sophisticated are only as good as the data they run on - a relevant well known saying here is “rubbish in, rubbish out”. Data quality has been a historical weakness of our industry, and this has been driven by the pace of change as well as a significant number of mergers without proper post-merger integration.

I believe that our focus needs to shift from use-case development toward cleaning our data and maintaining our data quality, so that pre-built, industry specific automations can sit on top of that data. Inventory data (especially physical) is very important, it is the keystone that much of our automation will depend on, and it is a historically weak area for most Telcos.

My further view is that our industry partners, such as the hyperscalers, will always be able to spend more on use-case R&D, building models we can then adopt, and this reinforces that our focus should be on clean, well structured data.

When focusing on data quality, it is important to consider this in two parts:

  1. The quality of existing data - this needs to be investigated, and improved where needed
  2. The processes which create and modify data - ensuring these are uplifted to stop the creation of poor data

Putting it all together, a feedback loop emerges:

Data feedback loop

My main parting messages are:

This is what I took away from the conference, and what I hope you will consider as a set of useful areas for focus in this coming year.

  • Collaboration across our industry is very important: no one organisation can match the research and development capacity of our collective group. We need to put aside competition as much as we can.
  • We have to continue to develop and embrace standards. This includes information models (like SID) and interfaces/APIs (like ODA). Through developing common ways of structuring and exchanging data, we can mix and match systems and models and enable greater levels of change at a greater pace
  • Data quality and completeness are paramount: and this is directly in our control, and one of the most impactful areas each of us can individually work on. Addressing this will be the main factor in allowing organisations to adopt common/shared models and automations.
  • North star: we have to develop our north-stars. These are clear articulations of the telcos we want to become including processes, products and technology, we cannot rely on evolution to get us there, and we cannot rely on use-case thinking.
  • Investment: this will all take investment, and we need to consider early and impactful changes we can make to strip cost from our organisations which will create investment capacity to allow us to keep making bigger changes

Here’s to a great 2025.

-- Richard, Jun 2024