A few years ago I worked directly for Liberty Global. Today I work for Infosys. The curious part is that I am still working, broadly speaking, inside the same technology ecosystem: similar platforms, the same markets, and many of the people I already worked with before.
What changed was the model from which we do that work.
In 2023, Liberty Global announced a strategic agreement with Infosys that included the transfer of more than 400 employees from areas such as Product, Technology Development, Service Delivery, Network & Shared Operations and Security. The deal was initially valued at about €1.5 billion over five years, and Liberty estimated savings of more than €100 million a year.
I was one of the people who lived through that change. And I want to be clear from the start: I am not writing this to judge whether one specific vendor is good or bad. I understand perfectly why a telco makes a decision like that. Scale, cost, specialization and access to capabilities that are expensive to keep entirely in-house. The question that stayed with me, looking back, is a different one: whether everything that is gained — and everything that is lost — really shows up in the numbers.
What changes is not only the logo on the signature
When the work sat inside a single structure, Product worked with country requirements, Epics were designed, we took part in planning and moved toward implementation. Of course there was bureaucracy. There were also discussions between countries, different priorities and projects that slipped. I do not want to paint a perfect organization, because it was not one.
The structural difference is something else. In a model with client, provider and several markets, the same decision often has to pass through more tables. When everyone is aligned, it works. When not, something that looks simple can start to move through conversations, escalations, commercial decisions and ownership lines. That is not necessarily anyone’s particular failure. Often it is the design of the arrangement.
And that friction has a cost: more coordination to keep the same pace.
The model also brings good things
It would be unfair to stop only there. In a contractual setup I also see more care around certain changes. There are KPIs and explicit responsibilities. Controls can be stricter. People think carefully before touching production because an incident has consequences that go beyond the technical team. That can help avoid problems.
In the relationship between Liberty Global and Infosys, improvements in some operational indicators have even been publicly reported. That matters. You can have better operational discipline and, at the same time, a different working dynamic. One thing does not cancel out the other.
You can have a better KPI and still feel more load day to day. You can have fewer incidents and move more slowly on new development. You can cut costs and make coordination more complex.
Efficiency is not measured only by looking at how much it costs to operate.
Doing more with less is part of almost every large transition
Teams usually change too. Some roles disappear. Others move offshore. People with many years of experience leave. And those who remain have to keep the platforms running, deliver projects and, at the same time, transfer knowledge and help train new teams.
That means that often fewer people end up covering more things. It is not exclusive to one specific deal. It happens in almost every large outsourcing transition. And that is where the word efficiency starts to look different depending on which side you look from.
For a company it can mean cutting costs. For a team it can mean more load. For a project it can mean less available capacity. And for a product it can mean that some things simply take longer, or that you no longer get the same level of attention you used to have.
That part usually does not appear when a major deal is announced. But it exists.
It is not only this case
Telcos have been outsourcing parts of their operations for many years. There are examples in Latin America too. Telefónica, for instance, agreed with Ericsson on network operations management in Colombia, Peru, Ecuador and Uruguay, including monitoring, incident management, problem management and change management.
It is not exactly the same model, but the logic has points in common: what do I really need to operate myself, and what can I hand to a third party?
That is a completely valid question. Keeping in-house every capability a modern telco needs is expensive. Networks, cloud, video, software, cybersecurity, data, automation. It is hard to be a specialist in everything. That is why outsourcing exists and will probably keep existing.
But I think there is another question that should be asked at the same time. What do I lose when I do it?
Because not everything can be replaced that quickly
In this kind of change, one of the hardest parts is watching people leave who knew the platforms very well. I do not want to overstate the point, because it is fairly simple.
If someone has worked with a system for many years and leaves, replacing their role can be relatively fast. Replacing their experience is not. You can document. You can run knowledge-transfer sessions. You can hire good engineers. But they need time.
Meanwhile, someone still has to keep operating. And usually that someone is the people who still know the system. That is why these changes do not simply happen on the day a contract is signed. They can take years to stabilize.
So, would I outsource?
Yes, some things. Selectively. And protecting critical knowledge above year-one savings.
There are functions where using external specialists makes a lot of sense. There are also moments when you need to scale quickly or access capabilities that do not make sense to build internally.
What I would avoid is getting to the point where practically all the real capacity to understand and operate a platform depends on third parties. Not because the vendor is bad. But because afterward every decision requires more coordination, and every knowledgeable person you lose costs more to replace.
At some point the savings start to carry a cost that is much harder to measure.
And now AI enters the picture
I think this is going to make the discussion even more interesting. Today a large share of technology outsourcing still depends on people. Often the equation is fairly simple: move work to organizations that can do it at greater scale or with different cost structures.
AI may change part of that equation. Maybe telcos themselves will be able to run certain functions with much smaller internal teams. Or maybe the opposite happens, and specialized providers appear that can operate platforms by combining engineers with AI in a much more efficient way.
We do not know yet. Today we still need people who understand the platforms, make decisions and respond when things really get complicated. But I find it hard to believe the current model will stay exactly the same.
Savings are easy to measure. The rest, not so much.
I think that is finally my biggest doubt about this model. It is relatively easy to measure how much a team costs before and after. You can also measure availability, incidents, SLAs or how long it takes to resolve a problem.
It is much harder to put a number on losing speed. On losing experience. On having overloaded teams during the transition. On needing more organizations to make a decision that used to need fewer. Or on product quality slowly ceasing to be exactly the same.
After living a change like this from the inside, I do not think the discussion is simply whether outsourcing is good or bad. I am also not measuring the performance of one specific vendor. I am asking what the business case should include when a telco outsources real operating capacity.
Because more efficient does not necessarily mean better.
You can shrink an organization quite a bit and still get it to keep working. The question is what happens afterward.
How much slower you move, how much knowledge you lose, and how much it costs to keep the same quality with fewer people.
Those things should also be part of the spreadsheet. Even if they are much harder to put there.
Related reading:
✍️ Claudio from ViaMind
“Dare to imagine, create and transform.”
FEATURED POSTS
ABOUT ME
ViaMind Ecosystem
- viamind.io · viamind.cl — The brand hub for ideas, technology and strategy.
- NeuraPRO — Practical software for managing real operations.
- ViaMind Radar — Less noise, more signal. AI-curated technology stories.