Leave a comment
Get the GH Bookmarklet

AMAs

Andy joined SoundCloud in February 2011 to help drive adoption of the audio platform's nascent mobile apps.  Over this time, he's seen the company change mindset to be truly 'mobile first', with mobile now accounting for the vast majority of growth and engagement on the platform and with mobile apps consistently in the Top 40 apps overall in the charts.  A little over a year ago, Andy set up a cross-functional growth team focused on user Retention - which he considers to be the most important growth metric - with a heavy focus on mobile.

Whilst at SoundCloud, Andy published his original 'Mobile Growth Stack' article and subsequently a 2015 update to the stack, based on his hands-on experience growing SoundCloud's mobile apps and his 16 years experience working in mobile.  This 'lightweight framework for strategic mobile growth' aims to help app developers to navigate the complex mobile growth ecosystem and identify opportunities for growth by taking a holistic approach.  The Mobile Growth Stack has been widely adopted as an essential tool for growth marketers and CMOs.

Andy started his career as a mobile game developer, having spent his childhood developing games for home computer systems.  He developed his first mobile title, a side-scrolling shoot-em-up arcade game called 'Space Impact', at Nokia in the UK back in 1999.  Space Impact was launched on the seminal 3310 phone and was subsequently embedded onto over 150 million mobile phones worldwide.  The game had many firsts on mobile, including fast scrolling graphics and use of vibration.  Andy spent much of the next decade developing mobile games before going to business school, specialising in marketing, and heading to Berlin to join SoundCloud.

The next step for Andy after SoundCloud is to continue to develop the Mobile Growth Stack and to help startups and larger firms to apply this framework in practice as a partner in the Phiture mobile growth consultancy, with fellow ex-SoundClouder, Moritz Daan.

Andy has a BSc Computer Science from Leeds Univeristy in the UK and an MBA from Warwick Business School. 

You can find him on twitter at @andy_carvell

He will be live on June 28 at 930 AM PT for one and a half hours during which he will answer as many questions as possible.

  • HQ

    Hila Qu

    almost 5 years ago #

    For anyone who haven't read Andy's Mobile growth stack post, be sure to check it out first:
    https://growthhackers.com/articles/2015-the-mobile-growth-stack-revised/#

    • AC

      Andy Carvell

      almost 5 years ago #

      Thanks, Hila :) Excited to be doing this AMA!

      I've also published the stack along with some tips on how to apply it, along with Moritz Daan (founding partner at Phiture and now co-author of the stack) at www.mobilegrowthstack.com - it's still a work-in-progress and we'll publish more stuff there over time, including updates to the stack as it evolves. Feedback welcome!

  • AW

    Abdelrahman Wahba

    almost 5 years ago #

    Hey Andy,

    Happy to see you here :)

    Here are my questions:
    1. What is the "North-star" analytics KPI at SoundCloud?
    2. What is the main retention driver at SoundCloud and what is SoundCloud's aha-moment?
    3. What is the average user lifetime on SoundCloud?

    I can understand you not being able to share the above info. If that's the case, would you please at least share how you identified the answers to the above?

    Thanks my friend :)

    • AC

      Andy Carvell

      almost 5 years ago #

      Hi Abdelrahman! Thanks... i'm honored to be asked to do an AMA on my favorite growth community!

      1. What is the "North-star" analytics KPI at SoundCloud?

      For SoundCloud, everything leads back to listening time.

      Even for creators; they upload their music & audio to SoundCloud to build an audience and be heard, so more listening time = more engaged listeners but also happier creators. It's also how creators get paid, through advertising and SoundCloud GO subscriptions (currently live in USA.

      2. What is the main retention driver at SoundCloud and what is SoundCloud's aha-moment?

      Ultimately SoundCloud is all about the content that creators upload to the platform; engagement (and hence retention) comes from the content. For listeners, that means we need to provide great ways to discover new content (improved search, charts, the new discover feature, etc.), listen for long 'lean back' sessions with recommended tracks and Stations, etc. Creators are interested in different things: they want to see stats on their tracks and get real-time updates when their music is going viral. That's why we built them a dedicated creator app, SoundCloud Pulse.

      3. What is the average user lifetime on SoundCloud?

      I'm afraid this is confidential. It's significantly different for listeners and creators, as you can imagine.

      (I'll answer some other questions and then come back to your 'how we identified them', but short version is that we spend a lot of time trying to figure out drivers of engagement & retention, both through analysis and experimentation)

      2 Share
      • AA

        Anuj Adhiya

        almost 5 years ago #

        re: "get real-time updates when their music is going viral"
        What patterns, if any, have you seen as to why certain pieces of music go viral vs others?

  • HQ

    Hila Qu

    almost 5 years ago #

    Hi Andy,

    Super excited for your AMA, I have been more focused on web product, but always found the world of mobile growth fascinating, so please bear with me if my questions are too amateur.

    1) What is your recommended essential tool stack for analytics/tracking, testing, user interaction etc. for mobile apps?

    2) When you talk about acquisition in mobile apps, does that in general mean # of installs? I would imagine there is a big fallout between awareness and install, and also delay in timing. How do you take that into consideration from a strategy perspective? And what are some tactics you can use to minimize the gap/delay?

    3) When you face a new mobile app, how do you think about which channels to try for growing this app? Do you have any framework?

    4) Lastly, how do you retain your users, if your user only need to use your app occasionally by nature (for example, home buying app, niche shopping app etc)? Personally, unless I fit this app into my daily/weekly schedule, or this app has a function that nothing else can replace, I tend to not use the app at all after 1st time usage. So I wonder what marketer/product designers can do to resolve that?

    • AC

      Andy Carvell

      almost 5 years ago #

      Hi Hila!

      1) What is your recommended essential tool stack for analytics/tracking, testing, user interaction etc. for mobile apps?

      I get asked this a lot! I try to remain somewhat impartial on the tools side, since there are so many vendors out there and lots of great tools and technology. However, I do make a point of checking out as many tools as possible and learning their pros & cons. I get to do this even more now since at Phiture we work with a large number of clients, all with different tech & tools stacks. I'll call out a few of the ones I see most commonly in use and enjoy working with, but I'd stop a little short of recommending a one-size-fits-all tools stack, because I don't think it's possible to generalize for the whole industry...

      For analytics, I see an increasing number of folks using Amplitude and I think they have done a really nice job with their interface. I've had great experiences with Localytics for analytics as well as app marketing.

      At SoundCloud we're currently using AppBoy for marketing automation and driving a lot of impact with it. I've also heard good things about Outbound.io for marketing automation. For analytics we have an in-house stack and event pipeline and use Redshift and Tableau for queries and dashboards. At SoundCloud we use Adjust for mobile attribution, Fabric in our mobile apps for crash reporting and a few other bits and are looking to use some parts of the Firebase platform. I also see a lot of clients using Branch or Yozio for deep-linking. We built our own A/B testing framework for in-app A/B but I heard good things about Apptimize and Swrve on that front.

      For ASO and App Store intel we use Priori, Sensortower, AppCodes, App Annie and a few other tools. There are tons!

      2) When you talk about acquisition in mobile apps, does that in general mean # of installs? I would imagine there is a big fallout between awareness and install, and also delay in timing. How do you take that into consideration from a strategy perspective? And what are some tactics you can use to minimize the gap/delay?

      Interesting question! Yeah, typically people usually consider mobile app acquisition to be driving the user to the point of install. You're right that the acquisition efforts, awareness and first touch might be before the install, unless the user is just browsing the App Store. Annoyingly, even an install isn't *really* acquisition, since a fair chunk of users don't actually run the app once they've installed it, so really an 'acquired' user is one that's seen in the app at least once (i.e. had >= 1 session). At that point, it becomes about activation & onboarding.

      Regarding the gap between awareness and install, it's complicated further for SoundCloud as we have widgets and a website. Most people come into contact with SoundCloud first through a player widget or integration in places like Twitter, Facebook and so on. Our acquisition team works hard to build direct, attributable and optimizable conversion flows from widget->app store and also mobile web->app store (i.e. cross-sell from our other properties). We also have a brand marketing team based in NYC who work on positioning, brand awareness, etc. For me, real 'growth' work starts when you can measure things, which is harder - though not impossible - to do with brand marketing, but it's undeniably important: just tougher to measure.

      3) When you face a new mobile app, how do you think about which channels to try for growing this app? Do you have any framework?

      I recently studied the SVBR growth course put together by Brian Balfour and Andrew Chen (awesome growth gurus!) and one of the modules was on exactly this topic. Brian talks about the 5Cs of channel decay and points out that by the time a channel is mainstream, it's already declining in value. I don't have a framework specifically around channels, but I think the closest would be something like the Growth Share Matrix (https://en.wikipedia.org/wiki/Growth%E2%80%93share_matrix) from classic business theory. It's important to maintain a portfolio approach and be investing in new channels while milking the 'cash cow' channels that are already working well for you (but ultimately will decline).

      4) Lastly, how do you retain your users, if your user only need to use your app occasionally by nature (for example, home buying app, niche shopping app etc)? Personally, unless I fit this app into my daily/weekly schedule, or this app has a function that nothing else can replace, I tend to not use the app at all after 1st time usage. So I wonder what marketer/product designers can do to resolve that?

      Retention is my favorite topic! Yep.. retention is very product-specific. A travel app might be super useful, but you only use it when you want to go on holiday. Hubspot arrived at 5-week retention as their KPI... it's important to understand what retention means for your app category. That said, platforms like LinkedIn have done a great job of building in additional engagement features to keep people coming back more often: content posts, endorsements, etc.

      You're right that the barrier to de-installing an app (or just leaving it dormant on the homescreen) is incredibly low. It's a very tough space. For me it all comes down to product-market fit: you need to be addressing a real need in the marketplace, but I disagree you have to be irreplaceable to succeed. A great - and ideally delightful - product experience and a very clear product proposition that's well explained to the user from the App Store onwards (into onboarding) goes a long way. Things like activity notifications and lifecycle marketing won't help you if users don't see inherent value in your product.

      • AC

        Andy Carvell

        almost 5 years ago #

        PS. I am a big fan in experimenting with new tools & services... multiplexers like MParticle and Segment make it easier to try out new tools and experiment with your stack.

      • HQ

        Hila Qu

        almost 5 years ago #

        "A great - and ideally delightful - product experience and a very clear product proposition that's well explained to the user from the App Store onwards (into onboarding) goes a long way. " Super insightful! Thanks Andy

        At GrowthHackers, we are implementing segment as well. The initial implementation has to be done carefully, but after that, it seems to allow much more flexibility.

  • AA

    Aldin A

    almost 5 years ago #

    Hi Andy,

    Thanks for doing this AMA.

    1)Can you talk about some of the challenges of scaling the sound cloud platform and how you've overcame them?

    2)How did sound cloud initially get the supply side of the platform when they didn't have the same credibility to attract creators? How did you build credibility with them?

    3)What was sound clouds strategy to get creators on the platform at scale?

    4) I noticed that you set up the cross functional growth team focused on retention. Can you talk about how you think about retention and improving retention? What factors do you look at when you figuring out how to grow retention?How do these factor differ on mobile vs. the web?

    5)How does retention differ between the supply side and demand side of the sound cloud platform?

    Thanks alot

    • AC

      Andy Carvell

      almost 5 years ago #

      Hi Aldin,

      1)Can you talk about some of the challenges of scaling the sound cloud platform and how you've overcame them?

      Scaling any two-sided marketplace is tricky, since both sets of users are critical to platform success, but often have very different needs. When you have small teams and limited resources, it can be hard to scale each at the same time. At SoundCloud we definitely focused more on supply side (creators) in the early days and then had a big shift towards growing the listener base, which is ultimately great for creators too. You've seen this shift in the product over the years as we've focused a lot more on building great listening experiences, but creators occasionally feel a bit left out when we release listner-centric features or things that make the product more accessible to the mass market, which is why we recently built a creator-centric app. You see this approach a lot; platforms such as Uber also have driver and rider apps, for example.

      2)How did sound cloud initially get the supply side of the platform when they didn't have the same credibility to attract creators? How did you build credibility with them?

      I wasn't there in the very early days, but I think the credibility of the founders, who were both into music and sound engineering and were 'scratching their own itch' when they built SoundCloud as a better alternative for music sharing and collaboration than posting CDs or uploading WAVs to FTP sites, went a long way. Also the decision to base the headquarters in Berlin - which is a hotbed of electronic music production and also home to great creator-focused tools companies such as Ableton and Native Instruments - was a smart move, I think.

      3)What was sound clouds strategy to get creators on the platform at scale?

      The company put a lot of work into building a creator-centric platform and also a community. The social features such as messaging, groups and particularly TIMED COMMENTS really helped to make SoundCloud a social sound platform and a true community. We also invested a lot in community support and engagement, through programs like SoundCloud Heroes, SoundClouder of the Day, etc.

      4) I noticed that you set up the cross functional growth team focused on retention. Can you talk about how you think about retention and improving retention? What factors do you look at when you figuring out how to grow retention?How do these factor differ on mobile vs. the web?

      Yep, our retention team at SoundCloud has growth engineers (backend, frontend, mobile), data science, design and growth marketing skillsets, as well as a PM. Growth is a team sport and isn't about building experiences so much as driving impact. We have a remit to touch any part of the product where we see a chance to increase retention, though ultimately retention is also a function of the overall product experience as well as our acquisition efforts (different channels bring different quality users) and the content itself.

      We look at listener retention from both a user perspective (users seen in the app) and a listener perspective (logged-in active listeners)... it's important to look at both, since one is an industry standard (user retention) and the other is filtered toward the engagement we are ultimately trying to achieve.

      We think of retention as a mixture of referring in existing users (through deeplinks), onboarding new users effectively (very important for any app and something where we know we can still do a lot better) and re-activating lapsed/churned users.

      In my team, we built out an activity notification service that listens to events on the platform and matches them to users who they are relevant for. This includes new content alerts but also social interactions. It's still a work in progress, but it's delivering great impact, mostly through push notifications, though we are in the process of adding email and web push as additional channels. It's a great capability that we can continue to build on... the next phase is about increasing relevancy and deeper personalization through machine learning and leveraging the usage data we collect to make smart bets on what users will enjoy being notified about. We also see great impact through lifecycle campaigns, especially using in-app messaging as a channel on mobile.

      Differences between web and mobile are largely the channels that are available - push is pretty killer on mobile and I'm excited to see if browser notifications can drive similar impact on web. To be honest, we focus our efforts on mobile, because we see inherently higher engagement there anyway; with limited resources, we'd rather put it into retaining mobile users, since they are more valuable in terms of engagement and ultimately more monetizable down the line.

      5)How does retention differ between the supply side and demand side of the sound cloud platform?

      The dynamics are very different. Creator retention is all about keeping creators interested in the platform and helping them build an audience... now with our Pulse app, creators can check their stats on the go, which is something they'd been after for a long time. Some of our tech stack can be repurposed for creators... activity notifications, for example.. we can use the same underlying delivery pipeline, experimental framework and analytics, but craft creator-centric notifications that only go to creators.

      • AA

        Aldin A

        almost 5 years ago #

        Thans so much for the informative answer.

        How do you go about figuring out the industry standard for retention in you particular industry?

      • AA

        Anuj Adhiya

        almost 5 years ago #

        Man I'm going to have to read this thread a few times - so much awesomeness in here.

      • AA

        Anuj Adhiya

        almost 5 years ago #

        I have to say the timed comments feature really is the most unique implementation I've seen - my mind was truly blown when I saw it the first time.

        Love the idea of the SoundCloud Heroes & SoundClouder of the Day - going to steal that for sure!

  • RB

    Ry B

    almost 5 years ago #

    Hey Andy,

    Great to have you here!

    1)What stops someone from recruiting the creators on sound cloud and creating a competing service - similar to what we've seen with Uber and Lyft vying for the same drivers? How does sound cloud 'lock-in' their creators? What is it that makes sound cloud a defensible product?

    2) What is the best advice you you received from someone? What's something you learned the hard way?

    3) Aside from retention what are key metrics you should be looking at in the early days of a platform startups?

    4) As a platform scales it becomes harder to match two parties that will like each other (ie the chance users will like the music they find on sound cloud). One way to deal with this would be a recommendation engine suggesting content the systems thinks the user will like. What has sound cloud done to make sure that users find content they like, at scale?

    Thanks

    • AC

      Andy Carvell

      almost 5 years ago #

      Hi Ry,

      1)What stops someone from recruiting the creators on sound cloud and creating a competing service - similar to what we've seen with Uber and Lyft vying for the same drivers? How does sound cloud 'lock-in' their creators? What is it that makes sound cloud a defensible product?

      Defensibility on the creator side comes through network effects & community as well as pure scale; other UGC platforms and copycat products are typically nowhere near as big as SoundCloud at this point and hence are less attractive to creators who want to reach a large audience. Of course, we can never be complacent there, and we do see some creators leaving SoundCloud for rival products, which is always sad. We hope they will come back and we continue to work hard to build a platform that allows creators to reach the biggest audience and monetize their work.

      2) What is the best advice you you received from someone? What's something you learned the hard way?

      Best advice was not to put all of our eggs in one basket when it comes to growth channels: SoundCloud were overly-reliant on the Facebook platform for a time - and it still contributes significantly to our growth - but we have diversified our referral channels considerably and continue to do so.

      Something we had to learn the hard way: Don't forget the V in MVP! At SoundCloud we rewrote our entire iOS app from scratch (for many reasons, but mostly to replace outdated buggy code with something that we could iterate on in a faster and more stable way) and knew we needed to get it out into the market as soon as it was 'ready'. We had to take some very tough decisions there in terms of cutting features from the launch version and - in my personal opinion - we released something that was a little short of viable from a feature-set perspective. Users loved the new design, but were taken aback by the decrease in functionality. Luckily, we were able to iterate on it quickly, thanks to the re-architected codebase and awesome development team & processes we'd developed over the past year.

      3) Aside from retention what are key metrics you should be looking at in the early days of a platform startups?

      Well, definitely retention ;) But it's also important to consider what engagements are valuable and think about High Value Actions (which are highly app specific but could be sharing, bookmarking, content consumption, purchases, ad views, etc.) and driving more of them. Finally, Days Seen (i.e. frequency of use on a per-user basis) can make a drastic difference to the stickiness and monetization potential of the product.

      4) As a platform scales it becomes harder to match two parties that will like each other (ie the chance users will like the music they find on sound cloud). One way to deal with this would be a recommendation engine suggesting content the systems thinks the user will like. What has sound cloud done to make sure that users find content they like, at scale?

      We're constantly working on this. Recently we've launched Stations (https://blog.soundcloud.com/2016/02/02/introducing-stations/), improved search and Suggested Tracks (https://blog.soundcloud.com/2016/06/22/introducing-suggested-tracks/)... these are all epic features for algorithm-enhanced discovery :)

  • SA

    Shaker A

    almost 5 years ago #

    Hi Andy,

    Thanks for being here!

    1) What skills does a marketer need to learn to 'understand' and 'analyze' data? Ex learn stats, sql, R languange, etc.

    2)There are a million things you could be working on growth at any one time. Can you talk about your process for
    figuring out the MOST important thing you have work on RIGHT for growth? How do you MAKE and PRIORITIZE your growth road map?

    3)What is your process for figuring out what channels you should focus on for customer acquisition? How do you evaluate them? How do you come up with a plan of action to attack that marketing channel?

    4)How do you think about scaling traffic growth? What changes for how you use a channel from when you first identify it's viability,
    to using the traffic channel at scale? What question do you ask yourself to make sure that the channels stays viable as you scale it up?
    What are the pitfalls to avoid as you scale a channel? How do you overcome said pitfalls?

    Thanks

    • AC

      Andy Carvell

      almost 5 years ago #

      Hi Shaker!

      1) What skills does a marketer need to learn to 'understand' and 'analyze' data? Ex learn stats, sql, R languange, etc.

      A good grounding in stats is definitely the place to start; understanding how to run experiments, generate an test hypotheses, statistical significance, etc. are really the cornerstones of scientific growth marketing and more important than the implementation specifics imo. After that, I think a general proficiency with SQL and spreadsheets will get you a long way. Beyond that, definitely being able to crunch data with tools such as R, pull data from APIs with CURL, Python or whatever, will make you more capable and more independent. I should point out that I'm still learning on this front. Luckily we have a great Insights team at SoundCloud that can help us out a lot with data analysis, but having these skills yourself is super-valuable, especially if you one day want to found your own startup.

      2)There are a million things you could be working on growth at any one time. Can you talk about your process for
      figuring out the MOST important thing you have work on RIGHT for growth? How do you MAKE and PRIORITIZE your growth road map?

      Yeah, choosing the right bets is super important. I actually just finished an article on this very topic over at http://www.mobilegrowthstack.com/applying-the-stack/ where I go into some detail on my approach and how we use the mobile growth stack at Phiture and SoundCloud as a strategic tool for opportunity identification and strategic planning.

      Ultimately, you have to consider growth holistically in order to prioritize effectively; there's no point scaling up acquisition if you can't retain or monetize the users you're acquiring. At a more nuanced level, you need to be able to quantify opportunities and do scenario planning. If you can't answer questions like "What will give me more active users over the next 6 months: increasing retention by 2% or acquisition by 5%, starting next month?", then it's very hard to make smart decisions on which part of the funnel to focus on and which initiatives to plan. So it comes down to having a decent forecasting model that allows scenario planning. Experience helps a bit when it comes to estimating impact, but you also need to be adaptable when things don't deliver the impact you hoped and know when to keep investing in making something better versus divesting / focusing on something else. The marketplace and technology also moves very fast, especially in mobile, so it doesn't always pay to plan too far ahead at the expense of agility. I'm a big fan of diversifying risk by investing in new capabilities that take longer to pay off, while also going for some quick-wins and ideally investing 20% of resources into 'moonshot' activities that have a low chance but very high potential upside.

      3)What is your process for figuring out what channels you should focus on for customer acquisition? How do you evaluate them? How do you come up with a plan of action to attack that marketing channel?

      Ultimately acquisition is about being where your target demographic is - ideally at scale - and approaching them there with a proposition/message that's engaging. For SoundCloud, our player widgets are great for acquisition, as is our mobile web player, which has terrible engagement (users typically bounce after listening to one track), but is a critical part of our acquisition funnel, since it attracts a ton of long-tail SEO traffic as well as referral traffic from social network and messaging app shares. Widget and mobile web experiences are a great 'try before you buy' experience a bit like a playable advert; our challenge is to convert these listeners into app users.

      When evaluating new channels, it's important to consider the scale... you need to go for really big networks if you want to acquire users in scalable way. Facebook has been great for that reason. You also need to consider the dynamics of the channel.. WhatsApp has hundreds of millions of users, but the sharing dynamics there (typically one-to-one or one-to-a-small-group) are radically different to Facebook or Twitter shares in terms of reach (lower on WhatsApp) and engagement (higher on WhatsApp, because messages are more direct and trusted). Ultimately you need to consider reach, conversion rate and the quality of the users the channel brings in. If it's a paid channel, then obviously price too :)

      4)How do you think about scaling traffic growth? What changes for how you use a channel from when you first identify it's viability,
      to using the traffic channel at scale? What question do you ask yourself to make sure that the channels stays viable as you scale it up?
      What are the pitfalls to avoid as you scale a channel? How do you overcome said pitfalls?

      Lots of questions there! As I've mentioned elsewhere in the AMA, it's important to maintain a portfolio approach with channels, since a great channel today might not be a great channel next month or next year. At SoundCloud most of our acquisition (and a significant chunk of re-engagment) comes from viral shares and SEO. We've put a lot of work into our integrations and sharing flows with social networks and a lot of ongoing work to opimize those flows as the channel dynamics evolve. For scaling viral content shares, we can scale it in several ways: by helping users find more 'sharable' content / driving more sharing activity, by plugging in new channels to share to / ways to share (although being mindful of cannibalisation effects when adding new share options) and by optimizing our CRO from 3rd party network back to SoundCloud. Depending on the channel, we don't always have so much control on the last part. As I mentioned above, it's also about tracking your user acquisition cohorts by acquisition source, so we can understand the value of each channel in terms of downstream engagement, monetization, viral potential of the users these channels bring us. Paid channels are more inherently scalable and we are now able to scale up our UA efforts in countries where SoundCloud GO is live :) But these channels have their own issues in terms of quality of users, fraud, etc. and also require us to have decent LTV estimates, which is tricky when we have a brand new subscription service that hasn't been out in the market very long.

  • SE

    Sean Ellis

    almost 5 years ago #

    Hi Andy, thanks for doing this AMA with us. I'm going to be in Berlin in October, so my question is very self serving. What other companies do you look at in Berlin to find inspiration about new ways to scale growth? Anyone you recommend that I should meet with while I'm there (people that are effectively scaling growth)? Hope we get a chance to connect!

    • AC

      Andy Carvell

      almost 5 years ago #

      Hi Sean,

      It's an honor and a pleasure :)

      Berlin has a thriving startup scene that's matured a lot since i arrived 5 years ago. Dubsmash are killing it lately, so I'd recommend meeting them for sure. I've also been really impressed with CaptnCook... they're doing some really cool stuff on the retention side of things ('disappearing' content that you have to save the same day - by coming back into the app - if you want to come back to it) and content marketing.

      Number26 are really innovating in the mobile banking space and are scaling very fast and EyeEm are doing great things and growing + monetizing sustainably.. awesome product too!

      At Phiture, we've been working with a few other Berlin startups such as OneFootball, who are doing some great event-based stuff around the European Championships and Clue - a period tracker app for women - which is seeing some great traction worldwide.

      If you have time, I'd love to grab a coffee / beer with you while you're over :)

      4 Share
      • SE

        Sean Ellis

        almost 5 years ago #

        Thanks Andy, awesome to have the inside scoop of who I should meet with! I'd definitely love to connect with you for a beer or coffee too.

        Thanks again for the fantastic AMA today!

  • LS

    Logan Stoneman

    almost 5 years ago #

    Hey Andy - thanks for your time today!

    I recently saw an article you shared about Facebook's work with artificial intelligence to become a better search engine. I was wondering what your thoughts are on artificial intelligence and development in the future? How do you think it will change the way users interact with apps and what are you most excited about in regards to these developments?

    Thanks!

    • AC

      Andy Carvell

      almost 5 years ago #

      Hi Logan,

      Yeah AI is going to change everything! Self-driving cars, automation of many service professions, etc. It's a fascinating field. I'm super interested in AI, bots and the whole space and am currently working as a growth advisor to an early-stage startup here in Berlin who have a recruitment bot called Job Pal that helps you find jobs via a simple chat interface: check it out! www.m.me/jobpal They're investing a lot in AI and NLP to improve the recruitment experience for recruiters and job seekers alike.

      Most bots are a long way from decent AI as they stand today, but they'll get there over time, I think. I think in terms of apps & mobile, the most useful applications of AI to date are Google Now (which is increasingly impressive) and to a lesser extent 'assistant' services like Siri, Amazon Echo, etc. and their chat-bot equivalents.

      In terms of where this is heading, I'm looking forward to far better recommendations based on a really deep understanding of my tastes and context, an end to form filling and other tedious processes and AI-generated content such as games, movies and music created entirely by the machine and personalized for me.

  • GD

    Guerric de Ternay

    almost 5 years ago #

    Hey Andy,

    Thanks for coming up with the Mobile Growth Stack. Very insightful.

    I have two questions regarding mobile content marketing:

    Today micro-moments (https://boostcompanies.com/micro-moments/) are getting shorter and shorter, i.e. people want they information they're looking for straight away. And technologies like Facebook Instant, Safari Reader View, Google AMP are making the mobile web looks the same everywhere.

    How do you differentiate on mobile?

    How do you convert on mobile?

    (in the context of content marketing).

    • AC

      Andy Carvell

      almost 5 years ago #

      Hi Guerric!

      Glad you find the growth stack helpful!

      I'm not an expert in content marketing so I will struggle to give you too much insight here. However, I agree that attention spans on mobile are very short :) It's important to bear this in mind, as well as the form factor, when designing content and also to consider the context of the reader; mobile content consumption tends to more atomic and favors multiple short interactions rather than long reads.

      CRO on mobile depends on what the conversion goal is... I think it's probably less useful to try to get a sign-up for a newsletter, for instance, though I see plenty of content marketing that still does this, so I guess it's working to some extent. If you're trying to convert to a payment, then supporting mobile-friendly payment methods is a pre-requisite. Converting to an app download is probably the most common... for this, you need to ensure continuity between the proposition you're presenting in the content with the way the product is positioned and described in the app store, since all of your traffic will land there.

      • GD

        Guerric de Ternay

        almost 5 years ago #

        Ok. I see your point. Thanks Andy!

        As you said, email marketing doesn't perform that well in mobile. We're certainly switching towards a model where retargeting matters a lot, as it makes the interaction seamless.

  • MM

    martín medina

    almost 5 years ago #

    Andy, thanks for coming on here and answering our questions it is great to have you.

    Retention is something that is talked about a lot in the mobile space. How do you think developers should approach retention and what is unique that you guys did at SoundCloud that helped you out with retention?

    A problem SoundCloud has faced in the eyes of many is regarding monetization. What do you think they should do to address monetization and how do you address monetization on mobile in your new role?

    • AC

      Andy Carvell

      almost 5 years ago #

      Hi Martín!

      Yeah retention is a huge topic! For mobile especially, onboarding is massively important; the majority of users will churn after their FIRST SESSION... this is incredibly scary. You literally have once chance to introduce your product, help users understand the value and get them to experience that value in their first session. If they don't, they are unlikely to ever use your app again :(

      We're looking a lot at our onboarding in the SoundCloud app as we know we can do a lot better. We've been running a lot of experiments with adaptive onboarding that helps users discover features we know they haven't tried, but we also are thinking a lot about contextual onboarding... getting smarter about what we know about the user from what they were doing BEFORE they downloaded the app and improving the relevance of their first experience through things like deferred deep linking and customized onboarding flows.

      I'd encourage lots of A/B testing with onboarding flows for new user cohorts... most of these users will churn anyway so you shouldn't be afraid to run bold experiments. I'm also a big fan of qualitative user testing and surveys to garner more insight into how users are perceiving the app; it's very easy for a team that lives & breathes a product to become completely separated from the reality of how new users perceive it.

      As for monetization at SoundCloud, we recently launched SoundCloud GO, our consumer subscription service with offline sync and access to a huge catalog of major label tracks alongside the 100+million tracks already on the platform. We sell these through In-app purchase or via web subscription. We also serve ads against certain content in regions where GO is live and are constantly iterating on new ad products (like vertical video) and building our roster of top-tier brands to work with.

      In my new role at Phiture, we work with clients monetizing in all sorts of ways: affiliate, subscription, advertising, freemium with in-app purchase, so we're getting a lot of hands-on experience helping companies with this. Things like bundling and retail merchandizing experiments, CRO, price differentiation / dynamic pricing and so on are all tools in the toolkit, but it's also important to consider your acquisition mix and acquire users that will monetize well.

      • MM

        martín medina

        almost 5 years ago #

        Andy thanks for your thoughts! I love the idea of contextual onboarding, this is now something I will be looking into in the future. Keep up the good work!

  • AL

    Arsene Lavaux

    almost 5 years ago #

    Hi Andy,

    Thanks for doing this AMA.

    1) What is your favorite mobile attribution platform? Why?

    2) What's your take on organic vs. paid growth for a mobile app? Based on your experience how would you stage each along the growth cycle?

    3) Where do you see the future of mobile growth?

    • AC

      Andy Carvell

      almost 5 years ago #

      Hi Aresene!

      1. At SoundCloud we use Adjust, who are based in Berlin, have an open-source SDK, are a Facebook Measurement Partner and play nicely with re-targeting platforms, etc. But at Phiture we work with clients who use other tools such as Tune, Appsflyer, etc. and to be honest they all seem to work pretty well. I think whichever attribution providers can help bring more transparency and protection with regards to fraud will do well in the coming year or two, since it's a big issue on mobile.

      2. Organic acquisition is great if you can cultivate it, either through linear channels such as content marketing programs, PR, etc. or - even better - through compound growth from viral loops. If you have great content in your app, as SoundCloud does, then virality (and also search) can be huge drivers of organic growth. A lot of companies overlook ASO keyword optimization which for me is short-sighted, since a lot of people search in the app stores and *everyone* should put effort into CRO on their app store page, since all traffic lands there.

      Paid acquisition is one of the truly scalable ways to grow, but I wouldn't recommend doing it at scale until you are monetizing and have an idea of what a user is worth to you. It can be useful even in the early days to acquire reasonable size cohorts to do A/B testing and other product optimization, even if you're not monetizing, but there needs to be an ROI if you're going to spend serious money on it. I think Apple's new paid app store search will be an important new paid channel to master. Ideally, your product has decent virality and a decent user LTV, in which case your effective CPI (when factoring in organics from viral acquisition) is lowered and you can really scale your paid UA.

      3. The future of mobile growth... huge topic! I think bots are very interesting and could open up new ways to acquire and engage users on other platforms. Paid app store search is another big one. Smarter notifications that allow deeper interaction right there in the notification tray also change to some extent how we think about what constitutes an 'app', as do deep-links and deep content indexing and retargeting... apps are no longer just entered by the 'front door' of the home screen, so they need to be designed with these other use cases in mind when considering ongoing engagement.

      • AL

        Arsene Lavaux

        almost 5 years ago #

        Thank you, Andy. Great answers. Cannot agree more with 2. And definitely concur on 3.

        Yes, well aware of Adjust and competitors for 1., wanted to have your opinion on this increasingly important aspect of mobile growth.

        Merci encore!

  • ES

    Edward Stephens

    almost 5 years ago #

    Hi Andy,

    You've had a really interesting background can't wait for this AMA.

    1) What do you think it is about Sound Cloud that caught both consumer and artist's attention? It seems to have meaningfully differentiated itself from Spotify and it seems to receive less hostility from artists over licensing issues.

    2) What does the future of SoundCloud look like and how can you incorporate the latest sweet of technologies e.g. AI to continue to delight users?

    3) What are the biggest retention challenges SoundCloud has faced and how did you solve them?

    4) How divergent do you think web and mobile experiences will become and what are the true limitations of mobile as you see them?

    • AC

      Andy Carvell

      almost 5 years ago #

      Hi Edward!

      1) What do you think it is about Sound Cloud that caught both consumer and artist's attention? It seems to have meaningfully differentiated itself from Spotify and it seems to receive less hostility from artists over licensing issues.

      SoundCloud is more than just a streaming service with the traditional iTunes catalog (though we have this too now, with GO)... it was built for creators as a place for them to share, collaborate with each other, and build an audience. Now, we're also helping creators make money from their work. Ultimately, SoundCloud differentiates through the breadth of content on the platform, from creators of all types and levels of maturity... from aspiring teenage singer-songwriters composing and recording tunes in their bedroom, all the way up to big-name artists like Snoop Dog, as well as podcasters, DJs, comedians, and anything else you can think of. As long as we can continue to provide upcoming artists as well as professionals with a platform that helps them build an audience and promote their work, the range of stuff you will find on SoundCloud will be so much richer than the other streaming services.

      2) What does the future of SoundCloud look like and how can you incorporate the latest sweet of technologies e.g. AI to continue to delight users?

      We're investing a lot in machine learning, mostly to build great recommendation services and long-listening features such as Stations and Discover. AI and machine learning can also help us in other places, too, like notifications... my vision for notifications is something that is personalized down to user level, not just in terms of the content of the notifications, but also the channels used, the frequency, time of sending, etc... all unique to the user and adaptive based on their behavior. Self-optimizing systems have so much potential to align user needs with business goals.

      3) What are the biggest retention challenges SoundCloud has faced and how did you solve them?

      I've said it in other answers and I'll say it again: ONBOARDING/ACTIVATION is the biggest challenge. For any app. It's disproportionately important and really tough to crack, especially for a mass-market product that needs to appeal to a broad base of potential users.

      I've mentioned a few ways that we're tackling the onboarding challenge, but we have a way to go yet.

      4) How divergent do you think web and mobile experiences will become and what are the true limitations of mobile as you see them?

      For SoundCloud, we have quite different features on web and mobile and we aren't prioritizing feature parity, so I expect them to remain divergent. I honestly don't see many limitations in mobile... screen sizes can be an issue and interaction options are potentially more limited, but both can be solved with creativity and actually create more intuitive, pleasurable experiences when done well.

  • AA

    Anuj Adhiya

    almost 5 years ago #

    Hey Andy - great to have you on!

    What's been your biggest learnings on App Store Optimization?
    Also, what resources would you recommend to folks that know nothing about the topic get started with it?

    • AC

      Andy Carvell

      almost 5 years ago #

      Hi Anuj,

      Biggest learning for ASO is that Conversion Rate is everything! CRO on the app store magnifies the impact of all your other acquisition efforts - including paid - and reduces your eCPI. A/B test all of your assets, but especially the first screenshot and the title. And to localize your keywords and app store assets. We actually built our own app store testing tool for internal use before TestNest and StoreMaven came along, but now we just use Google Play Experiments to test stuff and port the learnings across to iOS.

      Shamless plug: you can learn all about ASO at our new site! http://www.mobilegrowthstack.com/topics/acquisition/app-store-optimization/

      2 Share
  • JL

    Jane Last

    almost 5 years ago #

    Hey,

    If you would be launching a new product, what would be the first thing you focus on to drive growth?
    (Saying the product itself is working flawlessly)

    • AC

      Andy Carvell

      almost 5 years ago #

      Hi Jane :)

      I'd focus a hell of a lot on the onboarding experience; really nailing that is key to retention. Aside from that, I'd try to build for virality from the get-go... it's hard to generate artificially, but you should at least give it every chance of succeeding. And everything starts with analytics... if you're not measuring it, you can't improve it.

  • AA

    Anuj Adhiya

    almost 5 years ago #

    Last one:
    I know you did primarily mobile retention but did you have a sense of any patterns on desktop as to what converts lurkers (ie visitors who visit - maybe a lot but never sign up) on SoundCloud into ones that will actually sign up? Is there also an average # of visits/days before someone will convert from a visitor to a registered user?

    • AC

      Andy Carvell

      almost 5 years ago #

      Yeah, on desktop we have an anonymous experience and don't force users to create an account. What works well for conversion is placing most of the deeper functionality (e.g. liking, reposting, playlist creation) behind the registration wall (but surfacing these buttons in the anonymous experience) and doing a good job to sell the benefits of a (free) account when users try to perform these actions.

      I can't share specifics on avg number of visits/days to signup on desktop, but I can say that we see far higher engagement from logged-in users and it's a core objective for us to drive signups. Aside from giving us more channels to interact with users (e.g. email addresses), account creation has a lot of benefits for retention as users build up their music collection through likes and playlists.

      2 Share
  • AC

    Andy Carvell

    almost 5 years ago #

    Hi Guerric!

    Glad you find the growth stack helpful!

    I'm not an expert in content marketing so I will struggle to give you too much insight here. However, I agree that attention spans on mobile are very short :) It's important to bear this in mind, as well as the form factor, when designing content and also to consider the context of the reader; mobile content consumption tends to more atomic and favors multiple short interactions rather than long reads.

    CRO on mobile depends on what the conversion goal is... I think it's probably less useful to try to get a sign-up for a newsletter, for instance, though I see plenty of content marketing that still does this, so I guess it's working to some extent. If you're trying to convert to a payment, then supporting mobile-friendly payment methods is a pre-requisite. Converting to an app download is probably the most common... for this, you need to ensure continuity between the proposition you're presenting in the content with the way the product is positioned and described in the app store, since all of your traffic will land there.

  • PD

    Pranav Divakar

    almost 5 years ago #

    HI Andy,

    Thanks doing the AMA here!

    When you joined SoundCloud in Feb2011, What are the first few things which you did to drive adoption?

    • AC

      Andy Carvell

      almost 5 years ago #

      Hi Pranav,

      The first things I did were to get the Android app into more 3rd party app stores (low impact, as it turns out, though Amazon and Samsung Apps were worth it), cultivated direct relationships with Apple and Google, and pushed for proper mobile analytics in the apps.

  • HQ

    Hila Qu

    almost 5 years ago #

    Hi Andy,

    One more, when you do A/B testing on mobile apps, what are the key difference from doing testing on web products? What tool do you use to do testing?

    • AC

      Andy Carvell

      almost 5 years ago #

      Hi Hila,

      I'd say the major difference is that it can take longer to implement and run a test on mobile, since it typically requires more development, testing and deployment times before an experimental feature is ready for testing. We also work in release trains, so a test needs to be built, tested and shipped in a release train, then users need to upgrade to the new version and THEN you can test :) Web is much faster on that front as we can redeploy many times per day.

      We test a lot of marketing and in-app messaging stuff using AppBoy to get around this release-train issue, but we don't use it for full feature testing; it's not designed for that.

      We built our own in-house tool for managing experiments, which now works across Android, iOS, Web and Widget... it's pretty awesome!

  • DD

    David Duckworth

    almost 5 years ago #

    Hi Andy,

    Thanks for taking the time to chat with us.

    1) How do you spend your free time?
    2) What is different about what you guys are doing at Phiture and who have you worked with already?

    • AC

      Andy Carvell

      almost 5 years ago #

      Hi David!

      In my free time, I mostly enjoy the 5 Bs: Berlin, Berghain, Biking, Boating and Burgers. Berlin's epic techno scene is what keeps me here; the quality and diversity of techno in Berlin really is unparalleled, but I am also a big fan of checking out new gourmet burger joints; a new one springs up in the city almost every week lately. I like cooking at home too, especially English Sunday Roasts. I'm also training for a marathon this year.

      At Phiture (www.phiture.com), Moritz and I get to use all of the knowledge and experience we've gained working on growth at SoundCloud and developing the Mobile Growth Stack and help a broad range of companies with their growth challenges. It's great learning for us and we really enjoy working with a range of different companies at different stages of maturity and across a range of industries and app categories.

      We also get time to work on new projects such as the www.mobilegrowthstack.com site, which we will use to publish case studies and in-depth articles about growth, organized around the growth stack framework. We'll also have some guest posts on there from industry professionals and experts in specific fields.

      • HQ

        Hila Qu

        almost 5 years ago #

        5Bs, haha, that sounds like something a marketer would say :)

  • AA

    Anuj Adhiya

    almost 5 years ago #

    One More: In one of your responses above you mentioned that you built our own A/B testing framework.
    What's generally the trigger that tells you to go down this path vs just using an off the shelf solution?

    • AC

      Andy Carvell

      almost 5 years ago #

      Hi Anuj,

      Yeah I think that just depends what kind of scale you're at. Tbh, I would have been more than happy to use an off-the-shelf service (as we do for A/B testing on the marketing side of things with AppBoy), but when you have a lot of engineers at a company, the tendency is to build more of your own stuff rather than buy it in.

      The fact that we are a platform and have users who interact with us on Widgets, Web, Mobile apps and maybe also third party products like Sonos means that managing experimental groups and holdouts across the entire platform is much trickier / impossible with client-specific commercial solutions, so this was also a factor.

  • AC

    Andy Carvell

    almost 5 years ago #

    Aldin A: For industry retention benchmarks, check out the stuff Adjust put out recently: https://www.adjust.com/mobile-benchmarks-q3-2015/

SHARE
62
62