1. Final Project Report
Blink Traffic: Stevia Angesty Stanford University
Ian Christopher ENGR 245: The Lean
Michael Feng Launchpad
Richard Kidarsa Winter 2011
2. Ini$al
Product
Idea
Mobile
applica+on
providing
crowd-‐
Shanghai Lahore
sourced
real
+me
traffic
map
in
developing
countries
Bangkok Jakarta
3. Meet the Team
Stevia Angesty
Michael Feng
Richard Kidarsa
Ian Christopher
Background: Investment Banking and Private
Engineering, Business
Background: Computer Science
Engineering
Equity in Asia
Expertise: Indonesian connections, logistics
Expertise: Programming Indonesian connections
Programming,
experience
Expertise: Finance, Marketing
Role: Technical framework, server-side
Client application development, website
Role: Financials, Strategy
Liaising with customers and partners,
development
market research
5. What
we
did
(part
1)
• Interviewed
market
leader
in
traffic
informa+on
services
• Interviewed
8
corporate
customers
in
Indonesia
6. Inrix
Interview
Market
Leader
in
Traffic
Data
Services
Customer
Segments:
Public
Sector
Media
Automo$ve
Mobile
7. Pass/Fail
Criteria
• Companies
in
Indonesia
do
not
have
the
technology
to
detect
traffic
• Large
companies
in
Indonesia
distribute
at
least
200
corporate
Blackberries
to
their
employees
• At
least
2
companies
have
strong
interest
in
buying
traffic
applica+on
8. Customer
Segment:
Corporate
Interviewed
8
poten$al
business
sector
customers:
“We
have
our
own
proprietary
traffic
system”
“Traffic
is
driver’s
responsibility.
We
will
evaluate
the
fuel
usage”
“We
only
give
out
limited
number
of
corporate
Blackberries
to
employees”
“Tardiness
is
tolerable.
We
use
motorcycles
if
it
is
too
late” cTc
9. Customer
Segment:
Corporate
The Dispatchers:
Customer
Sounds cool. It’d be
Archetypes
more efficient than
relying on drivers.
Users
10. Customer
Segment:
Corporate
Customer
The Managers:
Archetypes
We’d love to have
the app if it works
Users Influencers
and
Recommenders
11. Customer
Segment:
Corporate
Influencers
and
Customer
Recommenders The Procurement Dept:
Our drivers know the
Archetypes
the traffic better
than you!
Users Economic Buyers,
Decision Makers
12. Customer
Segment:
Corporate
Influencers
and
Archetypes
Recommenders
The Drivers:
Are you questioning
our expertise?
International, local, large, and small
companies
Users Saboteurs Economic Buyers,
Decision Makers
13. Pass/Fail
Criteria
FAIL
• Companies
in
Indonesia
do
not
have
the
technology
to
detect
traffic
• Most
don’t
but
large,
sophis+cated
ones
do
• Large
companies
in
Indonesia
distribute
at
least
200
corporate
BlackBerries
to
their
employees
• Largest
mul+na+onal
bank
gives
out
approximately
10
• At
least
2
companies
have
strong
interest
in
buying
traffic
applica+on
• 0
out
of
the
8
companies
we
surveyed
16. What
we
did
(part
2)
• Interviewed
market
leader
in
consumer
traffic
applica+ons
(Waze)
• Interviewed
2
Indonesian
web
startups
• Surveyed
98
Indonesian
commuters
17. Waze
Interview
100% crowdsourced
traffic targeting
consumers
2,500,000
2,000,000 Recently raised $25
Users
1,500,000
1,000,000
million Series B
500,000 round
0
2009 (est) 2010 2011
• How did they grow? Social and gaming features
• What markets are they focusing on? North America
• What are the key challenges? Data costs, cultural differences
• How do they make money? They don’t…yet
18. Jagoan
Interview
*
Indonesian
Social
App
*
Partnered
with
Retail
Companies
User
friendly
interface
while
SLOW
User
looks
savvy
connec$on
Loyalty
19. Disdus
Interview
*
Indonesian
Groupon
*
Poten$al
partnership
for
adver$sing
Troublesome
Hard
to
payment
mone$ze
before
method
cri$cal
mass
20. B2C
Customer
Feedback
98
2½h
1h
$8
web
and
phone
spent
in
heavy
surveys
in
spent
in
car
per
traffic
(<6mph)
on
gasoline
per
Indonesia
day
per
day
day
21. B2C
Customer
Feedback
20%
of
reduc,on
in
heavy
traffic
would
save
a
person1:
Blink
Subscrip$on:
50h
$42
saved
per
year
gasoline
costs
per
year2
$1
/month
1.
Assumes
5
working
days
per
week
and
50
workweeks
per
year
2.
Assumes
gasoline
usage
is
30%
of
normal
usage
during
heavy
traffic
22. B2C
Customer
Feedback
120
52%
Plenty
of
$me
to
use
mobile
devices
while
web
and
phone
surveys
in
USE
DRIVERS
commu$ng
Indonesia
46%
1. Mo$vated
by
3-‐person
HOV
lanes
2. Alterna$ve:
pay
“carpool
CARPOOL
jockeys”
$2.50
per
trip
23. B2C
Customer
feedback1
Blackberry
leads
usage
but
iPhone/iPads
important
Conclusion:
Focus on Blackberry first and iPhone/Pad second
What mobile device do you use in the car?
45% 40%
40%
35%
30% 28%
25% 20%
20%
15%
10% 8%
4%
5%
0%
Blackberry iPhone iPad other None
smartphones
1. Data obtained from web and phone surveys of 120 potential customers
24. B2C
Customer
feedback1
What do you do with your Never
Rarely
Sometimes
Often
Always
>= Sometimes
smartphones in the car?
Work
39
17
33
6
17
50%
Gather traffic information
66
28
17
6
0
20%
Read news
39
22
22
28
0
46%
Browse internet
28
17
11
33
28
62%
Community and game mechanics are
Socialize with friends
Play games
11
33
6
17
28
39
28
11
44
17
86%
58%
critical to driving usage and virality
How often would you use these Never
Rarely
Sometimes
Often
Always
>= Sometimes
features of our application?
Report incidents
22
28
22
33
17
60%
Join a chat
33
28
44
17
0
50%
Play mini-games
44
28
28
22
0
41%
Ask and respond to questions
39
39
28
17
0
37%
Find your friends and followers
22
39
22
28
6
48%
Earn points for driving more
33
39
22
22
11
44%
1. Data obtained from web and phone surveys of 120 potential customers
26. What
we
did
(part
3)
• Assessed
market
size
• Tested
demand
crea+on
via
website
27. How
Big
is
the
Market?
1. Located
outside
the
U.S.,
Western
Europe,
and
Australia
2. Mobile
penetra+on
rate
*
popula+on
>=
4
million
3. Popula+on
density
>=
2500
per
sq
km
4. GDP
growth
rate
>=
5%
China:
Beijing, Shanghai,
Shenzhen
Total cities: 24 Mediterranean:
Istanbul
North Asia ex-China:
Taipei, Hong Kong,
Total mobile users: 203.7 million
Seoul
Latin America:
Mexico City Africa:
Cairo, Lagos
South Asia: SE Asia:
Mumbai, Delhi, Jakarta, Surabaya,
South America: Kolkata Bangkok, Singapore,
Rio de Janeiro, Sao Kuala Lumpur, Manila
Paulo, Bella Horizonte,
Lima, Santiago. Buenos
Aires
28. Market
Growth
Plan
Rest of
World
Rest of Asia • Brazil (Rio de
• India (Mumbai, Janeiro, Belo
Kolkata, Delhi) Horizonte
• China (Bejing, • Buenos Aires
Southeast
Shanghai, • Cairo
Asia: Shenzhen) • Istanbul
• Surabaya • Hong Kong • Lagos
• Bangkok • Taipei • Lima
Jakarta • Singapore • Seoul • Mexico City
• Bangkok • Santiago
• Kuala Lumpur
• Manila
Market 1 2 3 4
Cumulative number 1 6 15 25
of cities
Cumulative mobile 10.4 36.9 109.6 203.7
users (millions)
Detailed income statement and assumptions in Appendix
29. Success
Depends
on
Virality
>
Churn
Ra+o
of
early
stage
virality
rate
to
churn
rate
=
2.00x
30. Success
Depends
on
Virality
>
Churn
Ra+o
of
early
stage
virality
rate
to
churn
rate
=
1.50x
31. Success
Depends
on
Virality
>
Churn
Ra+o
of
early
stage
virality
rate
to
churn
rate
=
1.0x
32. Demand
crea$on
via
website
1
2
3
“Not
a
landing
page”
Doesn’t
show
the
No
Indonesian
version
product
33. Demand
crea$on
via
website
-‐
results
Clicks CTR
100 74 2.00% 1.56%
68 1.25%
50 22 1.00% 0.59%
0 0.00%
1 2 3 1 2 3
People need to use the product
CPC us to maximize learning click
for Conversions per
1.5 1.34 6.00%
4.05%
1 0.62 4.00% 2.94%
0.54
0.5 2.00%
0.00%
0 0.00%
1 2 3 1 2 3
35. What
we
did
(part
4)
• Developed
server
backbone
integrated
with
OSM
and
Hadoop
• Built
a
working
Blackberry
applica+on
• Iterated
based
on
user
feedback
• Talked
to
poten+al
partners
36. Server
Backbone
with
OSM
and
Hadoop
OSM
DATA
BLINK
DATA
MAP
MANAGER
TRAFFIC
USER/EVENT
DATA
MANAGER
Front
End
AGGREGATION
Client
38. First
release
issues
Device
cannot
Different
connect
to
provider
semng
internet
for
Indonesia
Downloading
Fix
website,
problems
Provide
instruc$ons
People
don’t
Pop
up
window
want
to
leave
to
force
people
feedback
to
give
feedback
39. Second
release
issues
Bad
GPS
Use
data-‐assisted
signal
GPS
Show
Users
dislike
instruc$ons
and
Pop-‐ups
pop-‐up
once
Users
have
old
Support
older
OS
Blackberry
OS
40. Second
release
feedback
result
User
wants
to
So
far
50
Want
traffic
data
see
their
friends
downloads
41. Third
release
Implemented
User
can
locate
20
downloads
chamng
others
on
map
since
release
3
days
ago
5
new
users
introduced
via
sharing
Feedback on 3rd release:
Privacy
Baoery
life
Implement
Adjust
server
privacy
toggle
ping
rate
42. Key
Partners
AGM
Market
share:
44%
mobile
Interested
in
partnering
if
phone
service
&
18%
Disdus
will
adver+se
Blink
Blink
can
provide
them
free
Large companies are hard to negotiate. Blink will
internet
mobile
phone
service
traffic
informa+on
in
their
website
focus more on partnering with other Startups
Currently
looking
for
mobile
applica+on
to
AGM
owns
trucking
fleet
Blink
will
adver+se
Disdus
partner
to
increase
internet
(>100
trips/week)
in
Java
in
the
Blackberry
service
usage.
Telkomsel
region
which
can
help
Blink
applica+on
wants
a
large
number
of
in
genera+ng
traffic
data
users
before
partnering
44. Epilogue:
What
we
learned
• MVP
has
to
iinclude
virality,
not
just
traffic
to
nclude
virality,
not
just
traffic
• Legal-‐ese
is
difficult
to
handle.
It
took
us
a
long
+me
to
understand
this
• Users
need
to
test
the
product
in
order
to
maximize
learning
• We
need
to
assume
tthat
users
are
ttechnically
lliterate
ssume
hat
users
are
echnically
i illiterate
• Simplicity
is
everything
from
the
user’s
perspec+ve
• Users
who
like
our
idea
are
not
the
same
as
early
adopters
/
promoters
• Need
to
balance
tech
planning
and
implementa+on
• Difficult
learning
curve
to
master
technologies
like
EC2
and
Apache
• Blackberry
is
not
developer-‐friendly
and
it
spreads
to
the
server
• Gelng
user
feedback
is
a
lot
harder
than
we
originally
thought
• Difficult
to
approach
customers
large
and
small
the
second
+me
around
• From
mindsets
to
trends
to
technical
literacy,
the
tech
landscape
is
different
in
parts
of
the
world.
• We
cannot
spend
too
much
+me
thinking
off
/
wri+ng
an
elegant
solu+on,
but
you
also
cannot
write
garbage
• Code
organiza+on
really
starts
to
become
more
and
more
important
as
you
LOC
grows.
By
5,000
LOC
it
will
be
crucial
to
have
a
well
organized
code
base.
• User
mindsets
and
ttechnological
environment
ddiffer
remendously
from
mmarket
to
market
indsets
and
echnological
environment
iffer
t tremendously
from
arket
to
market
45. Epilogue:
• Is
this
a
viable
business?
– Poten+al
to
solve
a
hair-‐on-‐fire
problem
for
a
huge
and
growing
market
– Small
capital
investment
needed
to
validate
business
model
– High
ROI
• Will
we
pursue
it
aqer
the
class?
– YES!
“You gotta be in front of the wave to catch it”
55. Appendix:
Key
Assump$ons
Growth stages Early Mid Late Plateau
Market penetration rate 0.0% 3.0% 10.0% 15.0%
Churn rate 50% 50% 20% 10%
Virality coefficient 0.75 0.60 0.20 0.10
Promotion % increase 10% 3% 2% Population growth
Employees per city 5 10 20 30
Rent per city 1,000 2,000 5,000 10,000
Revenues
Costs
Ad revenues start when users per city Monthly salary per employee
$1000
exceed
10,000
Setup costs per city
$100,000
Views per month per active user
67.75
Technology costs per 1000 us
Advertising eCPM
1.00
ers
$20
Fixed promotion costs per city
$10,000
Premium revenues start when users pe Annual promotion costs per a
r city exceed
200,000
ctive user
$1
Premium percentage of regular users
17%
Tax rate
25%
Premium pevenue per user per month
1.00
56. Appendix:
Markets
Rank Land area
Density
GDP
GDP
Mobile phones
Market
(Global City / Urban area
Country
Population
Density)
(in sqKm) (per sqKm) per capita growth rate per capita size
1
Mumbai
India
14,350,000
484
29,650
3,400
8.30%
63.22%
9,072,070
2
Kolkata
India
12,700,000
531
23,900
3,400
8.30%
63.22%
8,028,940
4
Lagos
Nigeria
13,400,000
738
18,150
2,400
6.80%
50.30%
6,740,200
5
Shenzhen
China
8,000,000
466
17,150
7,400
10.10%
62.80%
5,024,000
6
Seoul/Incheon
South Korea
17,500,000
1,049
16,700
30,200
6.10%
97.20%
17,010,000
7
Taipei
Taiwan
5,700,000
376
15,200
35,100
8.30%
100.00%
5,700,000
10
Shanghai
China
10,000,000
746
13,400
7,400
10.10%
62.80%
6,280,000
11
Lima
Peru
7,000,000
596
11,750
9,200
7.80%
95.50%
6,685,000
12
Beijing
China
8,614,000
748
11,500
7,400
10.10%
62.80%
5,409,592
13
Delhi
India
14,300,000
1,295
11,050
3,400
8.30%
63.22%
9,040,460
15
Manila
Philippines
14,750,000
1,399
10,550
3,500
6.70%
73.60%
10,856,000
17
Jakarta
Indonesia
14,250,000
1,360
10,500
4,300
6.00%
73.10%
10,416,750
21
Cairo
Egypt
12,200,000
1295
9,400
6,200
5.30%
76.80%
9,369,600
25
Sao Paulo
Brazil
17,700,000
1968
9,000
10,900
7.50%
100.00%
17,700,000
27
Mexico City
Mexico
17,400,000
2072
8,400
13,800
5.00%
79.80%
13,885,200
28
Santiago
Chile
5,425,000
648
8,400
15,500
5.30%
100.00%
5,425,000
29
Singapore
Singapore
4,000,000
479
8,350
62,200
14.60%
100.00%
4,000,000
32
Istanbul
Turkey
9,000,000
1166
7,700
12,300
7.30%
92.20%
8,298,000
35
Rio de Janeiro
Brazil
10,800,000
1580
6,850
10,900
7.50%
100.00%
10,800,000
37
Hong Kong
Hong Kong
7,100,000
1100
6,455
45,600
5.70%
100.00%
7,100,000
38
Bangkok
Thailand
6,500,000
1010
6,450
8,700
7.60%
81.00%
5,265,000
47
Buenos Aires
Argentina
11,200,000
2266
4,950
15,000
7.80%
100.00%
11,200,000
52
Belo Horizonte
Brazil
4,000,000
868
4,600
10,900
7.50%
100.00%
4,000,000
90
Kuala Lumpur
Malaysia
4,400,000
1606
2,750
14,700
7.10%
100.00%
4,400,000
Total
mobile
users:
203.7
million
58. Appendix:
Blink
Customer
Archetypes
Commuters in large developing market cities:
- Working professionals with cars
- Working professionals using public transport
- Students
- Stay-at-home partners (Tai-Tai’s)