Originally from my blog https://dbarnetson.wordpress.com/
What the Lifecycle of TV Anytime Says About the Broader Apps Economy
When we started work on our app TV Anytime in 2010, we based it on a set of assumptions about the apps market. While we had our share of startup drama and execution issues, the underlying problem is that the apps economy has gone bankrupt for most startups in the last 18 months.
Development Costs are Rising
- Complexity: Apps that you can build in a few weeks have largely been done; the opportunities that remain are more complex, reviewers and users have high expectations in UI and quality.
- Development: We had a very bad experience with an expensive local developer (MBD); and ultimately were forced to go offshore to India for development. We had cloud and PC development as part of our team; in retrospect we needed mobile development in house as well but with good iOS talent commanding $200/hr we could not find such a person.
- Android is much harder than iOS: Building an app that is compelling on both junker phones and superphones is very difficult and costly. Our Android development time was about 2x our iPhone time for similar features and as there is no “gold standard” to develop for in Android we left fixing platform specific bugs and crashes post launch.
Customer Acquisition Costs are rising
- 500,000 apps rises the noise level – app stores have no discovery tool other than their “top” lists; a new brand can no longer expect to make the top 25 without some significant spend on advertising or grey market rankings push; we made it up to position 86 the day after launch.
- Your app may “go viral” but you can’t plan on it: predictable (paid) customer acquisition costs for free apps are $0.25 – $0.75/trial through a variety of means: facebook likes, CPM/CPA advertising, typical PR firm results, grey market rankings buys.
- Facebook Virality Has Fallen: This matters a lot – virality (the number of “free” users you can get for each greenfield user you buy) is key as it is exponential. In 2009, people talked about 0.7–0.9 viral coefficients – meaning you get 2.3 to 9 “free” users for each user you pay for. The changes in the facebook platform have dramatically reduced virality – we saw virality so low it was not easily measurable – I would be surprised if many “apps” see viral coefficients above 0.2 – 0.4 – meaning you get only 0.25 – 0.6 free users for each paid user.
- Facebook Connect Is Ineffective: We chose to have users’ single sign-on to our app using facebook connect. This was a mistake. Three issues: (1) We saw a 70% balk rate at the facebook sign-on (2) Facebook changed their policy during our development and only gave us proxymail user email addresses; facebook’s proxyymail system does not work well (~60% delivery rate, 5 hour lag) so we basically couldn’t do email marketing (3) facebook breaks and changes their APIs faster than we can keep up consuming engineering resources.
- Churn rates are Increasing: Its inevitable – more apps = more churn; we planned on 20%/week churn for free users and were surprised to have 55%/week churn on a free app.
Monetization is Falling…like a rock
- Mobile CPM Rates Head to Zero: When we started in mid-2010, iPhone eCPMs were $10 – $15 with VCs merrily predicting they would go up 2x or more. Today, we have iPhone eCPM of $2.31 and an Android eCPM of $0.07 (in context that means that to get your $0.50 customer acquisition cost back you’d need to show the user 7,143 banners in a one week half life…. My sense is that iOS eCPMs are just lagging, they have a long ways to fall as very few brands are marketing through mobile display advertising and a lot mobile display advertising is just a pyramid scheme of new app developers paying established ones to try and get downloads.
- Android In-app Purchase Rates 20x Lower than iOS: We were pleasantly surprised by a 30% in-app purchase rate for iOS repeat users of our auto-renewing subscription; but disappointed only a 1.5% rate in Android. Why the difference? Two guesses: (1) Android users have been conditioned not to pay for anything and (2) Android in-app purchase is a mess; far from the one button execution in iOS.
Apple is getting harder to work with
- Apple Approval Hell: 7 Day approval cycle? Try 13 weeks…3 appeals, 4 new binaries.
- Apple Changes their Mind Arbitrarily: 6 weeks after our 13 week approval we tried to add a new in-app purchase; we were told that our existing in-app purchase was now “non-compliant” as Apple had changed their interpretation of their own rules and only “actual content” was allowed to use auto-renewing subscriptions; not catalogs. Ohh, and they’ll be applying this new interpretation retroactively in a few months so no option to just leave what we have in store.
- Apple bled us out: We planned on a two week approval cycle and a launch before thanksgiving; instead it took 3 months and we had to launch in December (don’t do that, by the way). We thus consumed our marketing budget on engineering and had no coverage left for paid acqusition post-launch.
The net of this is that the apps economy is bankrupt; if you’re thinking of building and launching an “app”, you missed the window.
If an “app” is a way to engage users as part of your broader strategy, then maybe – but do you really need an “app” to do this or can you do it through a mobile website, webapp or something cheaper and easier to develop?
Don Barnetson was the co-founder and head of sales & marketing at DDT Software, Inc; he is currently enjoying some downtime with his wife and new daughter.