Oil and Water: Mobile Apps and Fast Iteration

by Brent Hurley (blog: http://www.brenthurley.com/)

It goes without saying that data-driven iteration is a tigerblood of sorts for tech startups these days.  The ability to quickly and cheaply run and test multiple iterations of websites has helped companies accelerate towards the elusive product-market fit, as well as introduce data-supported enhancements and optimizations for more fully-baked products.  The doctrine has been to launch early, launch often, test, rinse, and repeat.  Getting your product out into the wild quickly enables real feedback from real users, to either confirm or deny initial product intuitions before significant investments in time and money are made.  On the web, this practice is pretty much a given.  Far better to launch a rough sketch of your product quickly than to sit in a vacuum and develop what you consider the “perfect” product only to find out the world doesn’t think your baby is as cute as you do.

However, today’s burgeoning mobile app marketplace demands new rules for engagement.

A Different Landscape

There are three main reasons why it is difficult to iterate quickly on mobile applications. 
  1. Approval Process – Particularly within the Apple app store, each application (and subsequent update) must be approved by the Apple gatekeepers before going live.  This delays feedback and causes developers to bunch improvements, which can mute specific causation links between each improvement and subsequent usage effect.
  2. A/B Testing – While there are a number of mobile analytic products available today that measure absolute usage (e.g. Flurry, Admob Analytics, etc), it is still difficult and clunky to run proper A/B tests.  For websites, A/B testing is seamless and invisible to users.  For mobile apps, a specific control group must be selected (which can lead to biased results), and each new iteration must be separately downloaded and installed.
  3. App Store Dynamics & UX – While both Android and Apple have app stores where users can search for, review and download apps, users predominantly download apps based on the Top 25 Most-Downloaded category.  In addition, the process of downloading applications, whether free or paid, presents a hurdle for adoption that’s significant when compared to the web where users can simply click on a link to try out a new site.  As such, users rely heavily on ratings and reviews before investing time to download and try new apps.


Taken together, the facets of friction inherent in the mobile app marketplace today suggest developers should not fail fast and fail often, but rather wait and launch a more refined product initially.  In particular, the exaggerated feedback loop for mobile apps can cause half-baked products to be D.O.A. if the app receives bad ratings from early adopters.  In response, many developers have found that launching first on the Android market (which doesn’t have a thorny approval process like Apple) is an effective alternative to quickly get feedback and refine their product before launching on Apple.  Although discovery-driven planning is more difficult within the mobile app world, it is not impossible.  You must adapt to the parameters of the situation.

Perhaps mobile apps and iteration are closer to oil and vinegar rather than oil and water – still hard to mix, but, when combined, can help produce universally tasty things.


Popular posts from this blog

Quiz Time 129

TCS IT Wiz 2013 Bhubaneswar Prelims

The 5 hour start-up: BrownBagBrain