Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ


Choose your language

InfoQ Homepage News The Spotify Model is No "Agile Nirvana"

The Spotify Model is No "Agile Nirvana"

Leia em Português

This item in japanese

At Spotify, management and the way the organization works support teams and agile practices by growing people. But Spotify isn’t an "Agile Nirvana"; it’s hard to reach high performance with teams that are constantly growing, changing, and splitting into new teams. Joakim Sundén, team and leadership coach at Spotify, advises to be bold with the vision but take small steps to get there.

Sundén spoke about things that don’t quite work at Spotify and how they are trying to solve them at Lean Agile Scotland 2017. InfoQ is covering this conference with Q&As, summaries, and articles.

InfoQ interviewed Sundén about how things that don’t go quite well impact the product, what engineering managers and tech leads do at Spotify on a daily basis and their contribution to the success of Spotify, what managers discuss in one-on-ones and how that helps their engineers, and how to make improvement happen.

InfoQ: For those who don’t know the "Spotify Model", can you briefly describe it?

Joakim Sundén:To be clear, we never coined the expression "Spotify Model"; we have just been talking to people about what we’re currently doing. I guess you could call it a model in the sense of "simplified description", but not as something to "follow or imitate", to use two different thesaurus definitions of the word. Talking to others outside of Spotify I understand it to be everything from just the organizational structure with squads, chapters, tribes and guilds, to all of the contents of the white paper published in 2012 plus the popular engineering culture videos>, all the way to the "Spotify Model" being a surface onto which they project their own ideas of "Agile Nirvana". If the latter is the case you’re definitely in for a surprise if you would start working at Spotify.

InfoQ: At Lean Agile Scotland you spoke about some of the things that don’t work well at Spotify. Can you give an example?

Sundén: Don’t get me wrong, I love Spotify, the culture and the way of working we’ve been able to cultivate here, but it’s far from an "Agile Nirvana". For instance, one thing that surprises a lot of people with solid agile experience is that engineering practices, e.g., TDD, pair programming, clean code and simple design, aren’t that common. It’s definitely the exception rather than the rule to see squads doing any of these frequently. In fact, many squads aren’t even that disciplined with common agile team practices. They might have planning meetings, but they don’t necessarily follow up how they perform against the plan or reflect on how they can improve on it. Teams using kanban rarely uses WIP limits, or if they do they mostly ignore them. I’m not saying that this description applies to all our teams; with about 150 teams I wouldn’t know, but this is from my observations having worked with dozens of teams over more than six years and from conversations with lots of other coaches and Spotifiers.

InfoQ: Doesn’t that impact the product?

Sundén: Do I think our product could be even better if we were doing more of the agile practices and if we were more disciplined with them? Yes, I do. There’s always room for improvement and I know first hand what difference these practices can make. And we are constantly trying to improve these things, which is one reason we’re hiring many agile coaches. You also have to remember that hyper-growth has been the normal operating mode at Spotify for many years. This is, of course, a strength in many ways but it also comes with challenges. It makes it hard to reach high-performance with teams that are constantly growing, changing, and splitting into new teams. On the other hand, I would like to think that our culture in part makes up for what we lack because of this. We try to not work on big projects and big releases, but instead trust autonomous teams to experiment and find great solutions to our strategic needs and objectives. Compared to many other places where I’ve been, management and the way the whole organization works actually support the teams and agile practices rather than being an impediment for them.

InfoQ: What do engineering managers and tech leads at Spotify do on a daily basis; what is their contribution to the success of Spotify?

Sundén: Their biggest and most important contribution is without a doubt growing people. Growing people meaning both recruitment of new people and development of the people already working here, and also growing them into the chaotic environment Spotify can be. Hyper growth, high autonomy, and a culture of trust, openness and transparency means lots of people constantly need help navigating and making sense of all the information, opportunities and decisions we face every day. Our chapter leads play a huge and important role in this, what I would call, continuous cultural onboarding. Most chapter leads meet weekly one-on-one with all their direct reports to support them and challenge them to grow. They become a fairly fixed and stable point helping you through your life as a Spotify employee.

InfoQ: What do they discuss in the one-on-ones? How does that help the engineers?

Sundén: Most of the time in one-on-ones is typically dedicated to the engineer’s needs, whatever they might be. From more formal things like establishing and follow up on your development plan and finding the right coaching and mentoring for that, to more informal conversations about what’s going on in your life right now. The one-on-one is also a good time and place for feedback and recognition. It can sometimes be about operational trouble shooting, helping the engineer to find out who to talk to or to understand how things work. Sometimes the chapter lead will have to bubble up impediments, but we mostly default to help the engineers help themselves, to develop their own capability to function well in the organization.

InfoQ: What advice do you have for coaches and other change agents who want to make improvement happen?

Sundén: Well, it depends… ;) In general I would advise trying to understand the problem before jumping into solutions. Is it even the right problem to be solving right now? Try to go beyond hand waving and opinions and find out what really is happening. A good way to start is to ask people to picture what their scenario would look like if everything was perfect. This puts them into a positive frame and helps focus on great outcomes. Once you’re sure you’re working on an improvement opportunity that’s worth your time, try small time-bound experiments that you actually follow through on. Use what you learn to come up with the next step. I’ve found that the combination of being bold with the vision but taking small steps to get there is a good combination.

Rate this Article