Deezer #Startup Lessons Episode 2: The Growth Hacks

This is a series of articles on the story of Deezer as I lived it (Deezer is a global music streaming service / competitor of Spotify). These articles will highlight key principles that have been theorized to help startups succeed. Please note that this is a retrospective analysis, we did not know about these principles at that time!

I was the 12th employee of Deezer in 2008, I built and led the mobile team there. Deezer apps are now used daily by millions of people worldwide.



Growth Hacking​

You might have already heard about growth hacking (GH), or maybe about some famous growth hacks. Like how Hotmail hacked their growth thanks to a simple message « PS : I love you — get your free email at Hotmail.com! », added automatically at the bottom of each and every email sent. Or how Airbnb tapped into Craigslist’s audience regarding apartment rentals, which sky rocketed their growth rate. Or how Dropbox grew dramatically thanks to their referral program. At the bottom of this article, you’ll find a link to my Growth Hacking slides that cover the subject quite extensively, and that include these examples (and a few more of them) with explanations and images.

If I had to define growth hacking in one sentence, it’d be: “Marketing that rely more on Tech & Psychology than on Budget”

​Indeed, to grow Deezer in the early days, we simply had no budget for marketing. But among the team there were some techies with a business and growth mindset (like me), and some marketers highly tuned to tech and data.
 Working together, we combined these skills:

Growth Hacking Skills

​​Still, if we succeeded sometimes, we’ve failed many times. And that’s an issue with every thing you’ll read about growth hacking: successes are praised, but no one tells you about the numerous failures that happened before finding something that works!

It looks like this:

​Also, we got lucky. We sometimes found things that worked with quite low effort / low hanging fruit. Things have become much much harder since… ​​

The Purple Cow

The book “Purple Cow”, written by Seth Godin / recognized as a top marketing guru worldwide

​Here’s why I think this book is a MUST READ for marketers and growth hackers:

  • It emphasizes the importance of Product (among the Ps of marketing), and even more a remarkable product (that’s the « Purple » thing)
  • Hence the ever growing importance of Product Management discipline
  • It highlights how advertising has become much less effective and saturated over time, and the book almost completely kills this statement inherited from the (now past) mass marketing era: ‘we need more advertising to be successful’
  • It explains really well how the success of a remarkable product doesn’t rely on advertising, but more on the fact that its power users (sneezers in the book) will greatly spread the word and become evangelists

​Oh, I offered the book to many peers, and some of my friends involved in marketing told me ‘I learned nothing, this is really well articulated but it’s just the marketing of today’. And indeed, the book explains really really well the marketing of today, and the success of Facebook (and the whole social networks era), success of iPhone, success of apps like Shazam or Snapchat, … Success of Pokemon Go is related too — remarkable game tied to an already remarkable brand, with perfect timing (young people previously exposed to the brand now all have a smartphone and some means to buy goodies in the game). As an example, let’s take the success of the GoPro. I’m a surfer (well, bodyboarder) and this is what happened a few years ago: there were 30 guys in the water on a nice surfing day. A guy comes in to the water with this tiny thing with a lens on his surfboard. And every other guy notices that tiny little thing with a lens (remarkable has nothing to do with size!). ‘Dude, what is that fu…ing thing on your board? Are you really able to film yourself in the water, surfing on the wave? Unbelievable!’ Then it came to Christmas or birthdays, and 15 out of the 30 guys buy their own GoPro! But despite being really well explained by the principles in the book, none of these successes are included as examples (though there are other really good ones). The reason is simple: the book was written in 2003!!! Pure genius.

There’s also this video of Seth giving a talk at Google and he answers a question from the Google Maps’ team. At that time it’s maybe the 7th map service globally. And it really seems they then just applied Seth’s answer as a guiding rule / vision. GMaps became number one by far within a few years. Hint: success did not rely on advertising budget. GMaps is still THE remarkable purple map service.

Purple Music (oh Purple Rain!)

So, back to Deezer. In the beginning, when you discovered the service, you quickly faced a great aha moment : ‘Ok, so there’s this website that is legal and allows me to listen to all existing music in the world… for free… and the more I search, the more I find great music tracks!’ So yes, Deezer was a purple music service. This « Aha moment » alone lead to HUGE word of mouth. The more Aha you create, the more « Hey bro, I’ve got to tell you about this amazing thing I’ve just discovered!!! » — simple. But Deezer was also highly remarkable because music is a social thing. Eg., in 2008, you visited a friend, you liked what he was listening to and you noticed it was from a website on his computer => he told you how great Deezer is. You saw a co-worker listening to music with headset on his computer, you asked him about this website. You went to a party and the computer screen was displaying Deezer and the amazing playlist that was going on. Remarkable. Oh, in fact, 30 other people also came to this party, and they all noticed. Some of them started to use it for their next party, where they invited 30 other people… Users flew in like Purple Rain. However, not all products are meant to be remarkable, nor can you create it artificially. As a marketer / growth hacker, as simple as it seems, working to find how people are already naturally discovering your product is really worth it. Regarding methodology, I think customer development, 1/1 user interviews and qualitative market research are your best bets. You might find golden nuggets that will then allow to optimize discoverability of your product, hence virality. Like Deezer playlists for parties.

PR = Press Relations​

Apart from the founders Jonathan Benassaya and Daniel Marhely who hired me, the first person I met at Deezer was Sophie Samama. She was (and she’s still) in charge of PR & Comm. The Deezer brand owe her a lot. The learning here is that a startup with quite small funds had hired a PR specialist as one of their very first employees (5th maybe?). Jonathan was highly focused on PR with lot of time and effort, and Sophie drove the PR execution at full steam.
 In the first years of Deezer, there were literally loads of articles every time we launched a new feature or moved a pixel on the website.

​​Note that it was not at all a pleasant experience for the team — being so much under the spotlight added a lot of pressure, and our failures (see previous episode) were also dramatically highligted. We learned to deal with it, as with huge negative feedbacks on forums anytime we changed something, when there was a bug or when service was unavailable. At least, it meant people really cared about Deezer. There’s a whole section in Purple Cow about that, ‘Don’t fear to be criticized’ — it’s pure genius and I wish we could have read it while we were suffering at Deezer. But PR did not bring users directly, at least not massively. It was more a brand building and product awareness effort. ​

PR = Page Rank​

Daniel on the other hand was a tech guy, and the « shining in PR » thing quite bothered him. No issue with that, there were many tech related issues he and his team had to focus on, and one thing that Daniel was really good at was SEO. As mentioned, thanks to press relations, we had many articles on media websites and great blogs that where linking to Deezer.com. For those new to SEO, these type of sites are called ‘high authority websites’. If they link to your website, it means to Google that your website is valuable too, and that it has some sort of authority too. To measure this authority, Google computes an index called ‘Page Rank’. The higher the Page Rank, the more likely it is for a website to appear in top search results for its related keywords.

So, having all major French media websites and blogs linking many times to the Deezer website dramatically increased our Page Rank. At one time, it even reached 8!

​And Daniel was really happy with that. Indeed, any Deezer page with proper built-in SEO for music related keywords would appear in the very top results. So the tech team started to index pages on Google for each and every content on Deezer: artists, albums, tracks, playlists, … Nowadays, Deezer indexes roughly 30 millions pages on Google!
 Result? In 2008/2009 in France, if you searched for anything related to music (« Madonna », « Purple Rain », « Achtung Baby »…), the Deezer page for related content was among the top 3, if not the first search result on Google.

So PR ², the 1st significant growth hack at Deezer, was about combining press relations with amazing SEO. Millions of users discovered Deezer by searching music related keywords on Google. This alone is what ‘killed’ our main competitors in France (Jiwa, Radioblog, Wormee…) over a quite short period of time. In fact, once the top ranks on every music related keywords were occupied by Deezer, it became quite impossible for competitors to rank over Deezer. Or they’d have to pay for Google Adwords where Deezer didn’t have to, and in that case it was simply like an unfair advantage.

NB: please note that SEO has changed a lot since what’s related here.

The Viral Loop​

At the very beginning of Deezer, we quickly discovered people were copy pasting the links to tracks or album pages to share them via email. Easing the sharing process by adding sharing buttons was just then an obvious thing to do. The following step (obvious again) was to allow sharing on Facebook and other social networks. More music got shared via Deezer, but it was not massive. Massive virality happened when people could link their Facebook accounts to Deezer, so that the music they liked got automatically shared without user action. This growth hack was used heavily in mobile gaming (by Zynga for instance), and at that time the Facebook feed was open and every post was visible (since a few years, the feed is filtered by an algorithm). As a consequence, every piece of music that was shared on Facebook by a user was kind of an ad for Deezer, displayed to all his friends. Friends that did not know Deezer discovered it (acquisition of new users), and friends that already knew Deezer came back on the website to listen to the shared track (retention). Since it was recommended by a friend, it was highly likely that the receiver would also like the shared music… and share it as well!

NB: please note that social media has also changed a lot since what’s related here. That’s something growth hackers have to deal with: most growth hacks expire over time!

Note that the viral loop still works via email or even word of mouth. It’s just less amplified. Also, remember virality can’t be created artificially at all. It only works if people are already sharing naturally. For startups, I advise to interview users to figure out if they talk about your product, what are their motives. Also dig for which words they use, for example by asking « if you had to tell a friend about us in a few sentence, what would you say? » Tip: these might be better words than yours! Then figure out what would be the best way (eg, channel) to share to others. Note that it might not necessarily be social media (email, SMS and IM transform better those days). So, in the case of Deezer, the right psychological question to answer is « Why do people share music? What are their motives? » My guesses: Self branding (being recognized as the music connoisseur among their friends), ‘music discovery as a gift’ and emotions. This would need more user interviews to get a broader view, though.

In short, virality is not a technical thing. It’s a psychological thing. Dig into the psychology.
 ​

Mobile Deep Linking

​Just before I left Deezer, we worked in 2010 on a technical thing we called « URL Handler ». Basically, we implemented a way for our mobile apps to also open the shared http links to Deezer.com / handle web urls. The idea behind this was to leverage this virality also for our mobile apps. A better name for this mechanism appeared since: Deep Linking. (all details in articles provided below) If the app is already installed, it opens directly on the linked content. If it is not yet installed, the app store page is opened to trigger download.

I think that the viral loop over deep linking alone triggered more than 10 millions of Deezer apps downloads, not counting the huge impact in retention.


More reads:​

  • My slides on Growth Hacking link
  • Growth Hacking for Apps link (in french / link)
  • Update on Mobile Deep Linking in 2016 link
  • Buy The Purple Cow link
  • Seth Godin talking @TED here link (talks about Purple Cow) and here link
  • Seth Godin talking at Google on ‘All Marketers Are Liars’ and helping GMaps team at the end link
  • More books for innovators link

Read also ​Deezer Startup Lessons Episode 1: Pivots & Product Market fit


Originally published at www.thinkmobile.fr.