Monday morning, end of June 2014, around seventy people slowly gather in one of the larger meeting rooms in a Swisscom office in Zurich. There is coffee, snacks, a sound system with a microphone and live translation into English.

It's the morning of Swisscom's first PI planning meeting for the first Agile Release Train (ART) within Swisscom - the TV ART.

The TV ART

The term "ART" has less to do with art in this case, but is the abbreviation for "Agile Release Train", a group of teams that work according to modern management methods, the so-called "agile" principles. These principles include a strong customer focus, interdisciplinary collaboration and risk reduction through continuous delivery of valuable results (making a room habitable instead of building the whole house first and learning from this for the next steps).

Blue TV (formerly Bluewin TV or Swisscom blue TV) is being developed at TV ART: Live TV, TV recordings, video on demand and much more from Swisscom, including TV Box hardware and apps.

The TV teams had already started working according to agile principles two years earlier. The development of Swisscom's new IPTV platform had proven to be far too unpredictable for traditional, plan-driven project processes. The speed of innovation was so high that we wanted to continuously adapt our plans without losing sight of the big goal.

Discovery of PI planning

Using these agile methods, we developed the TV product up to the official product launch in April 2014. The processes for this (How do we plan, develop and test in a continuous manner? How do we work together?) had been gradually developed.

And then we came across early versions of the Scaled Agile Framework, SAFe. A process model for agile working methods with many teams. I was fascinated: Scaled Agile had written down many of the organisational patterns and planning approaches in a structured way that we had also developed ourselves.

But SAFe had also taken many other aspects into account, formulated new principles and added the big idea: Gather everyone involved in a room so that they can create a rough plan for the next 10 to 12 weeks - together: an event called PI Planning.

When big room workshops were still foreign

So there we were in June 2014 - the new TV2.0 on the market, the teams scaling the platform, stabilising all parts and adding new functions.

I moderated these two days together with my team at the time. I remember that it was structured and slightly chaotic at the same time. We wrote down what we wanted to achieve over the next few weeks on sticky notes. We attached the sticky notes to pinboards that each team had in their team area. It was loud in the room because everyone was talking and planning together - especially across team boundaries.

The magic paper process

We created Swisscom's first so-called "programme board": the visualisation of larger delivery objects on the timeline - and the contributions of the individual teams.The programme board was a large paper wall that we filled with sticky notes. If there were dependencies, we connected the notes with woollen threads. This paper-based setup had its own stability issues.

And I remember the first time I felt the magic of PI Planning: having all the people and all the information together in one room. People who agreed on their upcoming collaboration with a handshake. The strategy team clarified priorities, helped to continuously adjust the scope of the planned work and made hundreds of decisions that would otherwise have required a lot of separate meetings.

Decentralised and collaborative planning

At the end of the first PI Planning, when we conducted the "Confidence Vote" by a show of hands ("How confident are you that we have created a good and realistic plan?"), I was relieved: we had a plan that everyone could agree on. Everyone had immediately agreed to do their best to realise this plan. Everyone had created the plan themselves - not just a project manager with Microsoft Project. That day, we added a lot of democracy to our planning process.

For almost ten quarters from mid-2014, PI Planning remained something that only the TV ART did. The next ARTs didn't start doing it until around 2016/2017.

The human connection

We invited suppliers to TV PI Planning early on. These partners had previously worked together on projects for many international telecommunications companies. It was only at Swisscom PI Planning that many of them met for the first time.

Over time, many Swiss companies attended the TV PI Planning. They wanted to understand how this collaborative planning process could also help them with their complex and dynamic projects. Many of them tried it out afterwards.

Benchmarking: Are we doing it well?

We also travelled around Europe to talk to other telecoms providers about our way of working.Although we were never able to carry out official comparative studies, we found out in conversations at the coffee machines that Swisscom was ahead in practically every aspect: customer satisfaction, time to market, (smaller) team size, market share, profit margin. And as a rule, we were very clearly ahead.

Our agile planning and working methods probably played a role in this.

Is PI planning on the rise?

Fast forward to 2024. Ten years after the first PI Planning, digital product development at Swisscom is still running in 10 to 12-week planning cycles with now over 50 ARTs.

At the end of each PI, the teams come together to review what they have achieved, to review and adapt the way they work and to plan the next cycle.

I am still amazed at the value that well-executed PI planning creates: the speed of decision-making, the achievement of a solid agreement on a common plan to which everyone has contributed.

Even though we should strive for the greatest possible independence from teams, PI Planning helps when multiple teams need to coordinate. You will achieve better results faster.

Not just a process

And of course there are also downsides.

I remember being invited to a "secret" PI Planning meeting once around 2018: The teams had just completed the "official" PI Planning and told me: "That was just theatre. We don't have any dependencies with the people in the other PI Planning. But nobody wants to hear that. That's why we're doing a second planning with the people we really need to coordinate with."Fortunately, we were able to resolve this somewhat crazy situation (sometimes we are also a strange large company) a short time later.

And just recently, a colleague told me that they never address the real problems in their planning because bad news is not accepted. They would just mechanically go through a few lists. Risks would go unnoticed. You can imagine the consequences.PI Planning is therefore not just a process that can be carried out mechanically "just like that". It requires a culture of trust, curiosity and commitment. It requires a healthy culture.I see this culture at Swisscom every day - with a few exceptions - and I am still enthusiastic about it.

Here's to the next ten years

I am very grateful that I had the opportunity to start collaborative big room planning at Swisscom ten years ago. I am grateful that our culture at TV ART allowed for honest and open discussion, even of problematic topics. It paid off, it was fun, and I wish the same for others in their working environment.

Will we still be doing PI Planning in ten years' time? Will we then be in a state where everything runs continuously and there are no more dependencies? AI agents that do all the work for us? Maybe. And maybe it will still make sense to get together every few months to tell a common story and make plans together.

Simon Berg

Simon Berg

Director of Engineering - Data, Analytics & AI

More getIT-articles

Ready  for  Swisscom

Find the job or career to suit you. A career where you can make a difference and continue your personal development.

What you do is who we are.

Go to careers

Go to current cyber security vacancies