Here’s a recap of Twitter’s #HelloWorld event Feb 2.

– add plugin to tons of third party services

– 18 billion crashers per month
– used on 1.3 billlion android devices
– non fatal iOS error loggings
– support for unity
– tvOS support

– show you live the events on you app
– 198 billion sessions per month
– support OS X and tvOS
– launch day experience: show summaries such as “successful release”, “low adoption”, etc

– phone number verification
– find friends using phone numbers
– use email verification to switch from email login to phone login
tvOS support
– voice fallback; if after a minute the sms did not go through, send a voice call

– ad exchange
– can add custom ads from your partners
– video ads

twitter kit
– add native video

– 1 billion unique visitors
– 40% smaller js library
– collect tweets by topic or user;
— curated collections have an id, which you can embed in any page to show the curated results

– twitter data
– see want people think of you product or related products
– Harvard used twitter data to track tweets about food poisoning to see places that make people sick
– each tweet has 65 fields; lots of data about the users

public search vs gnip
– gnip has better searches and filters
– gnip has replay to search streams back in time
– gnip searches all tweets back to 2006
– search every public as it is created

audience api
– pull statistics about a group of user ids. minimum of 10,000 ids
– can summarize interests, gender, etc of users


travel app powered by fabric and gnip

– list ten most tweeted attractions in a city
– show images and tweets about attractions
– show graphs of tweets per day



GA crash metrics were delayed several days
Twitter crashlytics can report crash within minutes

Ticketmaster mobile apps crash less than web apps during big on sale events because of the detailed crash data

disable new, buggy, slow features during large on sale events to reduce the number of crashes

use analytics to give dev team a score board
– add tracking to features so devs can know how users are using features they made

currently 57% success, 43% failure when users login;
– login failures are critical during large onsale events
– ticketmaster will simplify the log in process

Rollout new features to a select number of users
– use crash data to catch bugs before wide release
– use launch day experience data to make any adjustsments before wide release