Upload
greenrobot
View
1.507
Download
2
Embed Size (px)
DESCRIPTION
greenrobot’s latest app “Prime Guide” brings the TV program to Android Smartphones, Tablets, and Google TV. With a quickly growing installation count, great reviews and very high ratings in Google Play, Prime Guide successfully challenges big players of established brands. We want to present technical and non-technical aspects of app development: • The technical app anatomy • Android and server libraries we used (and wrote) for Prime Guide • Google App Engine as a backend: good and bad sides • Strategies on how to compete with “the big players” • The marketing campaigns we ran and their outcome • Successful customer support and presence in social networks • Stats and numbers http://www.primeguide.tv
Citation preview
Markus Junginger Android Developer Java Developer Open Source GDG Android Munich Geek & Tech Lead for
Prime GuideChief Android @
So, who‘s this guy?
Prime Guide – TV Guide App
Prime Guide: TV Overview
Prime Guide: Broadcast Details
Prime Guide: Gallery
Prime Guide: Tablet
Prime Guide: Google TV
Technology Overview
Prime GuideServer Cloud
Prime Guide App
SQLite Database
Remote ORM
Managers
GUI
Eve
nts
EPG Data Source ($$$)
Libraries used
Action Bar Sherlock Android Support Library EventBus Facebook Several internal greenrobot libraries:
Common Android stuff, bitmaps, cache, … In total: 12 client libraries used Libraries are essential for app dev!
Libraries written for Prime Guide
greenDAO Fast O/R Mapper (code generation) Object oriented access to SQLite Open Source
Protobuf RPC Binary Client/Server RPC protocol Based on Google Protobuf Multiple remote method calls per request
Backend – Why App Engine?
The 20:15 peak automatic scalability No OS maintenance / administration High availability Backed by Google and its infrastructure Good Java support Task Queues for background processing Easy deployment / version control
App Engine: nothing‘s perfect
No SQL (no, that‘s not cool)P.S.: Today there‘s Cloud SQL ($)
Proprietary APIs (no full JEE standard…) Data store impacted our data model
Entity groups, TX consistency, … Request time limits (no long polling etc.) Local server: slow, not 100% compatible Not the cheapest solution
Some Backend stats
Requests per second, 24h
20:15 New day, Widget updates
Some Backend stats
Number of serving instances, 24h
20:15 New day, Widget updates
Are we ready yet?
December 2012: Should we release? Personal impression of the app
About 100 TV channels are plenty Basics OK, some cool features still missing 100+ issues in our backlog, hmm…
OK, let’s release anyway… Before Xmas! Code crunch & last minute PR started Big question: Will people like it at all?
First day in the store
Beta Release on December, 21st
1,500 installs 4.6 stars rating Phew! Good start, how will it continue?
Can it keep the great rating?
Our last Minute Marketing…
Trying to spread the word We informed 1,000,000 blogs upfront One bigger mobile blog picked it up It was worth the effort
As expected, People liked…
Native Android look and feel Simplicity, thought-out UI Speed of the App First class tablet support Features
2D scrolling TV overview Integration of movie related web sites YouTube videos Reminders
And a few Surprises…
People were happy to give feedback Other features matter to them Most wished: add channel XYZ Top Missing features
Search Support for TV series (reminders, etc.)
Surprises continued…
Data collected during a 14-day period ending on March 4, 2013
Devices statistics from Google
Device statistics from Google
Just one problem with these stats:It shows all devices
The real question should be:Which devices install new apps?
Android Versions, Prime Guide
April, new data
Marketing Campaign (paid)
CPI campaign with a Android website that also run a 3rd party store
Good impact in 3rd party store:
Downloads over Time
Marketing Campaign / Play
Prime Guide featured on a big German Android web site plus newsletter coverage
So, what‘s the impact in Google Play? Unexpectedly low:
Downloads over Time
How to make people rate you
Ratings are essential for success Make it easy for users to rate your app But don‘t force it on them (dialogs, …) Our solution:
Info Screen
Great place to let users engage
Customer Support
Take feedback seriously Answer each mail personally (we try to) Value people‘s input, they invested time We track wishes and prioritize accordingly Implement feedback: makes people happy Takes time, about ½ - 1 day per week
Things underestimated…
Complexity of TV Data Legal issues with TV Data (Germany) Costs of TV Data App Engine quirks General complexity, both app and server
infrastructure
Optimize for Smartphones, Tablets and Google TV for initial launch.
Lessons learnt
One of our biggest mistakes?
Pick great new technologies, create them if non-existent.
Lessons learnt
Whatever, that’s the fun part.& Solid base to build upon.
Takes its time…
Do projects for living, too.
Lessons learnt
Focus and take risks.(If you can afford it.)
Release before you think it‘s perfect.
Lessons learnt
Reality check!Are you on the right track?
Release a beta version.
Lessons learnt
People are more tolerant.
How did we do since?
Rating improved by adding TV channels
~ 100.000 downloads Next milestone: leave beta with V1.0
New features Ads, or paid version
Disclaimer, Rechtliches
Alle Inhalte urheberrechtlich geschützt.
© Copyright 2013 Markus JungingerAll rights reserved.