Pressing Play on Bleeding Edge Products: What Spotify’s Autonomous Structure & Culture Can Teach the World

How does one company win over and continue to persuade people to pay for a product, where so many had failed before it? Looking at several key areas, we try to understand how aiming for maximum speed and minimal risk sets one business apart from the rest.

Spotify’s achievement can ultimately be attributed to their innovative approach to structure, culture and product management to future-proof the business. This includes a “minimum viable bureaucracy” where maximum autonomy and alignment are encouraged. It’s a system that in itself is fluid, agile and will continue to develop in the future.

‘Most Organizational Charts are an Illusion’ — Spotify Agile Coach Henrik Kniberg

From Snapchat to ‘Silicon Valley’, we are living in the age of the startup. We idolize their founders, aspire to their entrepreneurial values and daydream about working out of their cool offices. But what is it that’s so appealing about the startup? Is it the culture? The sense of adventure? Certainly there’s the excitement of the greater risks/rewards than one can find in more established companies. But is it also about making an impact on the world, of creating a ‘ding in the universe’ as Steve Jobs famously called it?

One such business, Spotify, launched in Sweden in 2008 and has been on an upward trajectory ever since. Essentially a music-on-demand app, it has grown to offer users access to more than 30 million songs, all for a $10 monthly fee. Currently adding ten million paying subscribers per year, it hit 30 million in total in March 2016. Set in the context of its main competition — Google, Apple and Amazon — this is even more impressive. And it leads us to a vital question: what is it about this music app in particular that makes it such a success? How does it win over and continue to persuade people to pay for music, where so many had failed before it?

In the following piece, we look into how Spotify’s structure, organization and approach to product development have helped it to outperform bigger, better-resourced rivals.

‘Waste Repellent Focus’

The first thing to note about Spotify’s internal structure is that it wasn’t always this way. The company started with a more traditional interpretation of Agile based around Scrum. Only later, as it grew, did it fine-tune this approach. Specifically, this was to future-proof itself against the bureaucracy and roadblocks it foresaw as it turned into a much bigger company.

As a movement, Agile started as a response to traditional project management: in particular, so called ‘waterfall’ methods and their high failure rate when it came to software projects. One way of seeing Agile is in moving away from a “throw it over the wall” approach, to technology working hand-in-hand with business. More specifically, it includes breaking down projects into smaller manageable, iterative parts.

As for the concept of lean, this emerged from the world of manufacturing, but caught on among startups with Eric Ries’ book ‘the Lean Startup’. We will delve into lean further on, with specific reference to product development — but suffice to say that it links with Agile, emphasizing a data and customer-centric approach, in which the idea of the ‘minimum viable product’ is central.

Autonomy, Mastery, Purpose

For the first step to understanding Spotify’s developer team structure and culture, look no further than Daniel Pink’s book ‘Drive.’

In this book, he lays out some considerably startling hypotheses about what makes people more motivated and productive at work. Backed up by some robust research, it’s fair to say his theories run contrary to years of established business thinking. He states that for any job involving some level of cognitive effort, performance is not improved by the promise of monetary gain. Instead, Pink proves that it’s actually improved by these three factors:

  • Autonomy — being engaged rather than compliant. That is, having the freedom to direct yourself, perhaps the opposite of micro-management.
  • Mastery — having the constant challenge and satisfaction of being able to improve at what you do. He cites the popularity of playing an instrument as an extra-curricular example of this.
  • Purpose — being able to make a positive change in the world. Back to Steve Jobs.

And to those who dismiss these ideas as too radical, or just misguided, he asks: why do so many people devote so much time to working for free to support services like Wikipedia and Linux?

Minimal Viable Bureaucracy

“A strong enough culture can get away with a volatile, informal structure.” — Spotify Agile Coach Henrik Kniberg

In short, Spotify realized early on that these motivational factors had to be baked into its organizational structure and culture. They would be key to the symbiotic aims of attracting and retaining the best engineering talent, and producing innovative, first class products.

The building blocks of this structure would be squads — multi-disciplinary teams, which are self-directed, according to a long-term mission. Crucially they also maintain the size, agility and feel of a small startup, usually less than eight people.

Scaling Agile at Spotify by Vlad Mysla

Squads are grouped into a wider matrix of tribes. Chapters, meanwhile, cover a specific function and include individual line managers. These exist independently from guilds, a loose, cross-team grouping which allows wider company knowledge sharing, and to which anyone is free to join or leave at any time.

Of course, the many different squads need to be tightly aligned, in order to fit in with the company’s overall targets and mission. Spotify therefore aims for “loosely coupled, tightly aligned” squads — that is, high autonomy — for fast development, coupled with high alignment. It sees this function something like this: Leaders define the problem, squads work out how to solve it.

“People are natural innovators, just get out their way and let them try things out.”

In terms of fostering that wider sense of alignment, community and innovation play their part. As an example, Spotify hosts a number of internal “unconferences” and encourages 10% of time to be spent on hack days — including full hack weeks twice a year.

In essence, all of the above initiatives speak to fostering that combination of autonomy, mastery and purpose, which Spotify believes are key to motivating high performing teams. The video above also hints at the company’s relentless focus on employee satisfaction: in one case a 91% employee satisfaction rate ‘wasn’t good enough’ and six months after addressing, had grown to 94%. A look at its company reviews on Glassdoor show some common complaints, of the type you might expect — disorganization at times, and duplication of effort — but still, a broadly positive approval rating of 4 out of 5 at the time of writing.

Minimum Lovable Product

We’ve looked at the structure that underpins Spotify’s engineering team. But what about its approach to product development: surely this, even more than any organizational chart, will be where a technology business flourishes or fails?

As mentioned above, at the heart of Spotify’s product strategy is the lean approach — and more specifically ‘think, build, ship, tweak’ — with an emphasis on building minimal viable product (or as they put it, minimum lovable product). This means having a quick release cycle for new products and moving onto the ‘tweak’ phase as swiftly as possible.

In fact, increasing speed and minimising risk are two of the company’s main considerations when it comes to managing this entire process. New features are first released to a small percentage of the user base, before further tweaking and releasing to a gradually wider group. In a clear tenet of both Agile and lean, features are then continuously updated, based on A/B testing and customer feedback data, right until they are retired.

Limited Blast Radius

Beyond the way development is handled, how Spotify’s products are built is another fascinating area. First, each part of the overall product (social, search etc.) is decentralized, immediately reducing the ‘blast radius’ should anything break. Beyond that, ‘feature toggles’ provide the same kind of logic on a micro level: as new features are developed within each product, they can easily be switched on or off without affecting anything else.

For completely new products, it always starts with the idea, which is scoped out and explained with the help of a story. This ‘data-driven purpose statement’ helps visualize how the product would fit within the overall Spotify brand. If approved, the process of building, shipping and continuous tweaking can begin. Just as it did with the Discover feature — which the company credits as one of its most successful releases to date.

Notably, Spotify’s CEO Daniel Ek is already talking up new features in development, including video, podcasts and weather. This fits right in place as the company’s structure and process are inseparably aligned to fast, risk-averse experiments leading in completely new directions.

Migrating Spotify’s Squad Model

Despite its apparent success, not to mention grounding in sound business research, most firms will see Spotify’s laser focus on autonomy as too great a risk to adopt themselves.

That said, its structure is not entirely dissimilar to other matrix organizations. As to whether the Spotify model can work elsewhere, we should look at some specific examples.

Trade Me, a New Zealand-based ecommerce startup similar to Ebay is perhaps the best example we have. Trade Me’s adoption of Spotify’s engineering team structure came about after they noticed a visible slowing down in development cycles:

“We continued to grow but adding new people no longer meant getting more done, if anything we were slowing down. Without realizing along the way, we had set up a web of dependencies… That meant what should have taken a week, could take a year and the harder we tried, the more complicated it got.”

Trade Me Chief Product Officer, Trent Mankelow

The response to this was to go all-in, forming Spotify-style squads, chapters and guilds:

The Self Selecting Organisation — Total Squadification at Trade Me by Sandy Mamoli

Chief Product Officer, Trent Mankelow, explains in a piece for CXO how despite a ‘nerve-racking’ setup process, the new model has overall been seen as a big success, and no-one would go back to how it was before.

For other firms that have drawn similar conclusions in their approach, look no further than Assembla. Using its own version of the feature toggle, it also religiously tests feature usage once live. Or for one of many startups that buys into the benefits of the hack day, Atlassian devotes one day per quarter to staff working on whatever they want.

The adoption of a Spotify-like model across a whole business, even beyond engineering, can be seen in Zappos’ ‘holacracy’ — a completely flat structure. Reports since its adoption have been mixed, with some complaining of chaos, or loopholes simply allowing new bureaucracies to replace old ones. Overall, it looks like an experiment in progress. And it may be a few more years before its overall success or failure is clear.

Towards a Flatter, More Autonomous Culture?

But what about the vast majority of larger, more established businesses? Where do they stand on the axes of alignment and autonomy?

One thing is for certain — we have yet to see with any great regularity a combination of flat networks or matrix organizations existing alongside (or inside, or even in opposition to) the traditional hierarchy and bureaucracy within large organizations. Will more companies look at adopting this model, even if only as an experiment in future?

Spotify’s structure is still fairly unique — fiercely democratic, relaxed and flat, in many ways a reflection of its origins in Sweden.

Given all we know, it’s perhaps surprising that Spotify actually tries to avoid recruiting people who are too dogmatic about Agile. In a way, this is great insight into the heart of its culture. It is committed to a system that it has personalized and adapted, and will continue to do so. That is, until that system becomes out of date. Perhaps this is the only way to truly future-proof a business?

While older, larger and more established companies may never be able to get on the cover of Wired, appear in ‘Silicon Valley’ or even practice the majority of what Spotify preaches, there are still answers here for all:

How to attract the top talent of the future. How to make your business more innovative, move faster, or recover more quickly from failures. There may even be answers here for how to be a market leader.

Or, maybe just to help you put a ‘ding in the universe’, as Steve Jobs so aptly put it.


This article is part of The Paragon of Innovation series by Amdocs Delivery exploring some of the most exciting achievements and developments in technology and engineering. In each case, we look for the innovation at the heart of all such great achievements.

For over 30 years Amdocs has accelerated business value for its customers by uniquely combining BSS, OSS, network control, optimization and network functions virtualization with professional and managed services. A global company with revenue of $3.6 billion in fiscal 2015, Amdocs has over 24,000 employees and serves customers in more than 90 countries.

Amdocs: Embrace Challenge, Experience Success.

For more information, visit Amdocs at www.amdocs.com


Disclaimer: The information provided in this article is provided “as is” with no guarantees as to its veracity. You’re welcome to challenge us, disagree with us or just comment below.