Revision on "What I Did to Launch my iOS App"

Remixing the launch plan - among other things

 Hitting you with the remixed launch plan

Hitting you with the remixed launch plan

Growth is very nuanced. There is a unique intersection between marketing, product and data where this field really gets interesting. It requires both creative and analytical capabilities to really succeed. My journey into this field began when cofounding my startup Dribble. In an early attempt to document this journey I wrote about how I launched my iOS app.

As mentioned in the previous post, the essay depicted the launch strategy I used, as well as being used as a content marketing campaign. Since hindsight is 20/20, I thought it best to look at both the launch strategy and the campaign in one go. 

I plan to tackle this revision by looking at the bigger campaign-picture, in terms of how I set the campaign goals, the user hypothesis, how i constructed the conversion funnel and the top-line performance, but also critique what parts of the launch strategy I'd adapt. 


Since budget was £0 one can't expect crazy numbers - unless you've constructed an intuitive, natural viral loop in your product (which will be covered in another blog post in future), or have incredible WOM, network effects etc. The majority of organic strategies take a lot of time and don't product huge amounts of users in the short term. They require time to trial, test, optimise and iterate. Since this was my first campaign I decided on a conservative goal. 

Conservative Goal: 100 registered users

Extreme Goal: 1,000 registerded users

Budget: £0

Conservative web traffic goal: 1,000 unique visitors


Hypothesis: through initial analysis of our userbase from launch, decided to target tech savvy entrepreneurial pros who tend to be early adopters and reddit readers. Keep segment relatively broad by reaching US, male, 24+ . US based due to their familiararity of the daily fantasy game (the product I cofounded) - the assumption that the target audience will be the "lowest haging fruit".


Believed that the natural funnel for a content marketing campaign would hit landing first, bypassing the App Store. 

Conversion Funnel

KPI:  10% conversion rate from 1,000 web visitors / Tracking: using Branch to analyse attribution




  • Generated more web traffic than expected


  • Registered users fell short
  • Conversion rate fell short

Where in the funnel would I dig deeper to analyse/optimise/iterate:

I'd start by analysing each stage of the conversion funnel to find which stage had the biggest drop-off

  • If "posting content" - I'd try new mediums/content channels to maximise initial traffic
  • If "landing page" - I'd test which copy, creative, CTA each content medium preferred to maximise click through
  • If "App Store" - run a weekly experiment to multivariate test creative and copy to maximise conversion
  • If "internal metrics" - would run experiment to analyse churn, drop-off at each stage of onboarding and registration flow. Can run test on different "a-ha" moments (if not found already) to funnel different cohorts of users to different "a-ha" moments and test each cohorts' stickiness



1. Research research research.

I'd supplement the existing tools and growth leaders with these individuals:

  • Brian Balfour  - cofounder of Reforge
  • Nir Eyal - best selling author for "Hooked"
  • Justin Mares - author of "traction"
  • Gabriel Weinberg - author of "traction"
  • Sean Ellis - CEO, Growth Hackers
  • Brianne Kimmel - Growth Marketing, Zendesk
  • Fareed Mosavat - Senior PM Growth, Slack
  • Neil Patel - Cofounder, KISSmetrics
  • Eric Peters and this - Senior Growth Marketing Manager, HubSpot Academy
  • Rebecca Rosenfelt - Product Manager, Airbnb
  • Alex Schultz - VP Growth, Facebook

2. Social.

Identify which social channel(s) most naturally fit with your product. I initially thought spreading the social real estate across almost every social channel would be best, it turns out specialising is the way forward (big surprise...). Because it not only allows you to focus, but makes life a lot easier (segmenting which channel focuses on acquisition, customer service etc). 

Here's how I started:

  • Twitter - conversation
  • Instagram - imagery
  • Facebook - everything
  • Reddit - football chat
  • Slack - experiment
  • Quora - thought leader

Here's what worked: 

3. Metrics

Choose your software wisely. I mentioned the use of Mixpanel and Intercom. Both are amazing products, however, if you don't have the luxury of a high burn rate (which we did), we had to change to Amplitude (free to use - except for behavioural cohorting). It has the same tracking and analytics as Mixpanel, although with a few more features for its free version. I also believe it has a more intuitive design and layout. Used for all retention curve analysis, onboarding flows and "a-ha" moment ID.  

Since our data was on MongoDB I couldn't use SQL (a transferable skill I was interested to learn), so we elected to use another powerful database querying tool, Metabase. This allowed me to create scheduled queries based around key events in the football season, and manually look into behavioural cohorting. 

4. Inbound Marketing

Over the course of Dribble's lifetime, I have tried many different content strategies. From Dribble's own blog, social-only infographics to a standalone blog called Verge90. Key is to find the one which speaks to your users the most and be patient and consistent. Since I didn't have the man-power to build out a stellar content strategy (talking consistent content, guest posting, writing for many publications) as well as taking on all other growth channels, I had to make the painful decision to focus only on social-only. Ideal scenario would to build content assets for the business, around Verge90 (a standalone thought leading blog). 

5. PR

Choosing this channel depends on what your strategy goals are. If it's awareness, investor interest or general fluff, then PR is great. There's a whole strategy I will write about which includes my email outreach template, press kit and press release. It helped me get into tech Crunch, The Times, The Guardian and Huffington Post. If you're thinking of using PR for user acquisition, then best not go for it. Focus on more tracklable channels.

6. Try new things

The Slack group didn't work. But I'm glad we tried it. 

Other trials:

  • Across the Euro's - built a landing page called (site is now down). Experiment was to acquire cheap footbal focused users who are looking to stream the Euro games online. Note: this didn't convert as many users as expected from the traffic, main reason is that the Euro games were free on TV on select stations. 
  • Send a friend a beer (courtesy of Seffa)- sponsored christmas campaign to send people a tin of beer for the holiday
  • Gave out flyers at football matches in London (Arsenal, Chelsea etc)
  • FREE BEER - if you download and register on Dribble - targeted stadium traffic outside big stadiums
  • Scraped emails from large Slack groups and conducted email marketing campaigns

7. Networking


Product Hunt

Meetups - create your own or join some

Wework events

Coworking events


8. Paid advertising

Great way to test:

  • Targeting capabilites
  • User profiles / hypothesese / segments
  • Onboarding flow
  • Retargeting and acquisition
  • Funnel testing (cold acquisition, lookalike audiences, Facebook Pixel targeting)

It's multivariate and AB testing on a budget - test your £100 on any of the above. If that works you can scale it. If not, then try another test

You can use:

  • Facebook
  • Twitter
  • LinkedIn
  • Adwords
  • App Store

9. ASO

What I covered previously still works. Just as you would focus on improving SEO on your website, do the same for your app. Keyword analysis is key, comp analysis etc. App Annie will help. 

To summarise, I went through how I set the campaign goals, the user hypothesis, how i constructed the conversion funnel, how the campaign performed, but added to parts of the launch strategy.

If you missed "what I did to launch my iOS app" you can find it linked.