8. A Few Of Our Customers
25B
Messages Sent
50K
Applications
500M
App Downloads
9. Push messaging is complex
Brands
Platforms
Consumers
Cellular Wifi Fixed Line
Urban Airship:
Complexity Creates Scaling Difficulties - Connects all platforms and devices
- Compatibility with many platforms & devices - Rich Push & Location
- Push is the foundation for mobile marketing - Massive Scalability
- Data, Scalability, Location & Context
- Ease of Use - provides one easy interface
- Any network, Cellular, Wifi
10. Basic Push
•
Easy
Integra,on
-‐
bundled
libraries
adds
messaging
to
your
app
quickly
•
Scalable
Pla4orm
-‐
One
API
communicates
to
all
pla:orms
•
Push
Composer
-‐
Put
the
power
of
push
in
the
hands
of
marketers
•
Audience
Targe,ng
-‐
target
and
personalize
the
message
•
Advanced
Repor,ng
-‐
Get
relevant
insights
to
improve
and
opAmize
11. Rich Push: Create Immersive Experiences
• Call
to
Ac,on
-‐
Invite
users
to
take
acAon
in
the
app
• Persistent
Inbox
-‐
Users
won’t
miss
your
messages
• Stylish
Experience
-‐
create
branded
experiences
12. Powerful tools for marketers
Push Composer Reports
Measure:
• Pushes
• App
Opens
• Time
In
App
• Opt-‐in/Opt-‐out
13. Push Composer
See a preview of what the
user will see as you draft
your Push messages
14. Measure Success With Rich Reporting
Metrics: App opens, Time in
app, Pushes to app
User view across Push, Rich
Push, In App Purchase, and
Subscriptions
· Download Rate
· App Open Rate
· Time In App
· App Usage Statistics
15. Driving Critical Business Metrics
• 20% increase in total mobile orders
• 540% increase in daily app opens
• 3X faster response time than email
• 30% increase in social sharing
(on Facebook & Twitter)
17. Location and Context Targeting
Send a push notification based on location, history, preferences, and behavior
Create custom targeting zones: Target specific locations:
• City, County, or State/Province • Stadiums
• Zip, Neighborhood, DMA • Parks
• Congressional, School, and Voting Districts • Airports
• Census Tract • Beaches
• Country • Store locations
• Time Zone • Customer/user data
18. “All noise is waste. So cultivate quietness in your
speech, in your thoughts, in your emotions. Speak
habitually low. Wait for attention and then you low words
will be charged with dynamite.”
-Elbert Hubbard
20. The power to “light-up a stadium a second”
• High-throughput messaging - 75K messages per second
• Scalable connection handling - millions of concurrent connections
• Lightning-fast, low-latency, delivery
• Light-weight network footprint - low-bandwidth, low-power usage
21. The power to “light-up a stadium a second”
• High-throughput messaging - 75K messages per second
• Scalable connection handling - millions of concurrent connections
• Lightning-fast, low-latency, delivery
• Light-weight network footprint - low-bandwidth, low-power usage
22. Segments Use Case #1
• A news publisher could increase engagement by sending a
call-to-action to a "Big City Finance" segment
• Tagged with 'finance' or 'business' and located in Chicago
or New York City
23. Segments Use Case #2
• A retail manager could promote an end-of-season sandals
sale to a "Marina Shoe Shoppers" segment
• Tagged with 'product:shoes' and located in ZIP code
94123 within the past 30 days
24. Segments Use Case #3
• A sports app could conduct a cross-promotion with a free
beer offer to a "Giants fans attending a home game in
person" segment
• Tagged with 'team:SFGiants' and located within AT&T Park
tomorrow at 8pm
25. Segments Use Case #4
• An events app could remind potential concert attendees to
buy tickets by creating a "Beyonce fans in the Bay Area"
segment
• Tagged with 'artist_like:Beyonce' and located in 'San
Francisco Metropolitan Area'
26. Segments Use Case #5
• A TV network could introduce a new mystery show that
premiers in 1 hour to a "West Coast mystery fans" segment
• Tagged with 'genre:mystery' and located in the PST
timezone in the past 3 days
28. Location + Tags = Relevant Context
2.5M location boundaries (geofences) included
• Full historical support
“Send to users that were in this area in the past 30 days”
• Tags = Preferences and Behavior
• Combine location and tags
• A real-time interface to engage mobile users
31. How it works - Mobile SDKs
• Platform dependent
• Be smart with battery
• Reasonable defaults (significant-change service on iOS)
• Ability to use GPS and background collection if necessary
• Buffer until another network request is necessary
• iOS and Android SDKs ready to go
urbanairship.com/docs
32. How it works - Polygon Boundaries
• What is a location?
• 2.5M location boundaries (geofences) indexed in an R-tree
• City, County, or State/Province • Stadiums
• Zip, Neighborhood, DMA • Parks
• Congressional, School, and Voting Districts • Airports
• Census Tract • Beaches
• Country • Store locations
• Time Zone • Customer/user data
• Redwood converts latitude/longitude to meaningful boundaries
33. How it works - Bucket Indexes
• One a latitude/longitude is converted to ~12 polygon
boundaries, we update a ton of indexes in a kd-tree
• Polygon boundary by app by minute
• Polygon boundary by app by hour
• Polygon boundary by app by day “Show me all devices with app 1234 that
• Polygon boundary by app by week were in seen in Chicago in the past 30 days”
• Polygon boundary by app by month
• Polygon boundary by app by year
• kd-tree is like a B-tree, but you split on various dimensions
• Fully distributed on Cassandra, can handle high volumes
34. How it works - Querying
• Location and tag back-ends support sorting
• “GooeyButterCake” merge-joins results from each back-end
• Conditional logic using and, or, and not
• Mediator - responsible for producing the final set of device
IDs to deliver the message to
• More back-end data sources in the future!