The secret to building cross platform Mobile Apps: How the big guys do it - Droidcon Paris 2014

Post on 13-Jun-2015

209 views 0 download

Tags:

description

http://fr.droidcon.com/2014/agenda/detail?title=+The+secret+to+building+cross+platform+Mobile+Apps%3A+How+the+big+guys+do+it Today's mobile ecosystem is silo-ed. It's often hard to build for each complex platform, which are each important in their own right, but if we take a step back, the value of going cross-platform is the seamless experience you can build for your customers across the many devices they use throughout the day. Learn how Ali Parr, head of Partner Engineering at Facebook & Parse, has been working with the most disruptive tech companies for the last couple of years, helping them build cross-platform Mobile Apps that scale to millions of users and becoming highly profitable businesses. Speaker: Ali Parr, Parse (Facebook Company) Head of Mobile Platform Partnerships Engineering, EMEA, Facebook and Parse, based in London. He currently heads up the Parse partnerships program for Facebook in EMEA, as well focusing on new developer acquisition. Ali is focused on building partnerships between Parse and developers across EMEA, through direct contact, public speaking opportunities, and mentoring events. Ali is also a mentor at Techstars in London, and The Family in Paris, providing advice and experience to high-potential startups. Prior to Facebook, Ali was the founder of Infinite Degree, a gaming startup that reached top 10 in Apple App Store across many regions including the US and the UK. Ali holds a Masters degree in Computer Science.

Transcript of The secret to building cross platform Mobile Apps: How the big guys do it - Droidcon Paris 2014

ALI PARR PLATFORM PARTNERSHIPS, FACEBOOK, EMEA

ACCELERATING GROWTH BY DEVELOPING CROSS PLATFORM

1903First Flight

1903First Flight

1928 Jet Power

1903First Flight

1928 Jet Power

1947 Aeroflot

1903First Flight

1928 Jet Power

1947 Aeroflot

1969 Moon!

1903First Flight

1928 Jet Power

1947 Aeroflot

1969 Moon!

1973 Concorde

1903First Flight

1928 Jet Power

1947 Aeroflot

1969 Moon!

1973 Concorde

25 years 19 years 15 years 7 years

1903First Flight

1928 Jet Power

1947 Aeroflot

1969 Moon!

1973 Concorde

1903First Flight

1928 Jet Power

1947 Aeroflot

1969 Moon!

1973 Concorde

1903First Flight

1928 Jet Power

1947 Aeroflot

1969 Moon!

1973 Concorde

1960sARPANET

1903First Flight

1928 Jet Power

1947 Aeroflot

1969 Moon!

1973 Concorde

1960sARPANET

1993 Mosaic (Netscape)

1903First Flight

1928 Jet Power

1947 Aeroflot

1969 Moon!

1973 Concorde

1960sARPANET

1993 Mosaic (Netscape)

1999 .com Boom

1903First Flight

1928 Jet Power

1947 Aeroflot

1969 Moon!

1973 Concorde

1960sARPANET

1993 Mosaic (Netscape)

1999 .com Boom

2007 - 2008 iPhone / Android

1903First Flight

1928 Jet Power

1947 Aeroflot

1969 Moon!

1973 Concorde

1960sARPANET

1993 Mosaic (Netscape)

1999 .com Boom

2007 - 2008 iPhone / Android

2010 > Cloud

1. AS A TECHNOLOGY MATURES, SPEED INCREASINGLY MATTERS

$ git-clone-www ~/www

0

2800

5600

8400

11200

14000

2007 2008 2009 2010 2011 2012 2013

COMMITS / ENGINEERS PER MONTH

2. SUCCESSFUL PROJECTS SCALE

A STORY…

0

14

28

42

56

70

Q1-12 Q2-12 Q3-12

CANDY CRUSH CANVAS MAU

Q4-12

0

14

28

42

56

70

Q1-12 Q2-12 Q3-12 Q1-13

CANDY CRUSH CANVAS MAU (EXPECTED)

Q4-12

0

100

200

300

400

500

Q1-12 Q2-12 Q3-12 Q1-13 Q2-13 Q3-13 Q4-13

CANDY CRUSH MAU (ACTUAL)

Q4-12

THE WORLD WENT MOBILE

2005

2013

THE ENGAGEMENT LOOP

‘THE LOOP’ IN CANDY CRUSH• Progress synced across platforms - Levels, scores etc

• ‘Lives’ mechanic - false scarcity

• Friends and progress (via Facebook) - competition

• Sharing and stories - exposure

• Regular, small gameplay sessions builds loyalty (*cough* addiction)

About 0.15% of mobile gamers contribute.. 50% of all of the in-app purchases generated in free-to-play games.

Keeping these users engaged is vital!

About 0.15% of mobile gamers contribute.. 50% of all of the in-app purchases generated in free-to-play games.

IT’S NOT JUST TRUE FOR GAMING….

3. CROSS-PLATFORM ROCKS

FACEBOOK MOBILE USAGE

iOS Android Mobile Web

1.07b MAP

FACEBOOK MOBILE USAGE

2009

“Why has the Facebook app been so slow? Because the current version of the app is nothing more than a web browser inside an Objective-C shell. Stuff is

being constantly pulled from the web, hence the lag time.

The next Facebook app will replace all that sluggish sludge by using more Objective-C, which in turn would use more of the iPhone's hardware so it has to

collect less information from the Web."

- Gizmodo, May 2011

“Our biggest mistake was betting too much on HTML5."

- Mark Zuckerberg, September 2012

ENGINEERING FIRST

PEOPLE FIRST

4. CROSS-PLATFORM IS REALLY HARD

1. SPEED MATTERS 2. SUCCESSFUL PROJECTS SCALE

3. CROSS-PLATFORM ROCKS 4. CROSS-PLATFORM IS REALLY HARD

ALWAYS BE SHIPPING

Developer Workflow

Trunk www(SVN)

Developer Workflow

Trunk www(SVN)

Developer Workflow

Trunk www(SVN)

Local Branch(Mercurial)

Branch from Trunk

Developer Workflow

Trunk www(SVN)

Local Branch(Mercurial)

Branch from Trunk

Locally commit your changes as you go

Developer Workflow

Trunk www(SVN)

Local Branch(Mercurial)

Rebase regularly

Branch from Trunk

Locally commit your changes as you go

Developer Workflow

Trunk www(SVN)

Local Branch(Mercurial)

Rebase regularly

Branch from Trunk

Locally commit your changes as you go

Tests!

Developer Workflow

Trunk www(SVN)

Local Branch(Mercurial)

Rebase regularly

Branch from Trunk

Differential

Submit for code review

Locally commit your changes as you go

Tests!

Developer Workflow

Trunk www(SVN)

Local Branch(Mercurial)

Rebase regularly

Branch from Trunk

Differential

Submit for code review

Locally commit your changes as you go

Reviewer says yeah!

Tests!

Developer Workflow

Trunk www(SVN)

Local Branch(Mercurial)

Rebase regularly

Branch from Trunk

Differential

Submit for code review

Ready for commit

Locally commit your changes as you go

Reviewer says yeah!

Tests!

Developer Workflow

Trunk www(SVN)

Local Branch(Mercurial)

Rebase regularly

Branch from Trunk

Rebase again

Differential

Submit for code review

Ready for commit

Locally commit your changes as you go

Reviewer says yeah!

Tests!

Developer Workflow

Trunk www(SVN)

Local Branch(Mercurial)

Rebase regularly

Branch from Trunk

CommitRebase again

Differential

Submit for code review

Ready for commit

Locally commit your changes as you go

Reviewer says yeah!

Tests!

Tests!

.org

SHIPPING CAN BE PAINFUL…

…BUT IT LEADS TO UNDERSTANDING

MEASURE: TO LEARN ABOUT YOUR USERS

OKCUPID: LOVE IS BLIND

OKCUPID: LOVE IS BLIND

OKCUPID: LOVE IS BLIND

OKCUPID: LOVE IS BLIND

CREDIT CARD ENTRY EXPERIENCE VERSION BVERSION A

CREDIT CARD ENTRY EXPERIENCE

+2.4% lift in completionsPeople like their keyboards. They also like things that match

VERSION BVERSION A

VERSION A

DEBIT CARDS AND CREDIT CARDSVERSION BVERSION A

Credit/Debit CardCredit Card

+1.5% lift in rate of credit cards addedMake options obvious, even if it seems obvious to you

DEBIT CARDS AND CREDIT CARDSVERSION BVERSION A

VERSION B

Credit/Debit CardCredit Card

MINIMIZING CANCEL BUTTONSVERSION BVERSION A

MINIMIZING CANCEL BUTTONS

Conversion up but refund rate increased by 30%Bad experiences can raise conversion temporarily, but not permanently

VERSION BVERSION A

MIX

COUPON FLOW EXPERIENCEVERSION BVERSION A

COUPON FLOW EXPERIENCEVERSION BVERSION A

+30% lift in payer rateDon’t distract your customers from the main objectiveVERSION B

MEASURE: ENSURE YOU MEASURE THE RIGHT THINGS

LET’S PLAY A GUESSING GAME

ARPPU

% of payers

C

B

A

D

WHICH ONE HAS THE “BEST” MONETIZATION?

ARPPU

% of payers

C

B

A

D

IT ALL BEGINS WITH ENGAGEMENT

ARPPU

% of payers

C

OUTSOURCE, OUTSOURCE, OUTSOURCE

1. ALWAYS BE SHIPPING 2.MEASURE, ITERATE, IMPROVE

3. OUTSOURCE

THANKS, QUESTIONS?