The document describes the team, journey, product, customers, market, and business model of SET (Streamlined Embedded Technologies). SET optimizes sensor networks by providing middleware and reusable software subsystems. This results in more energy efficient, reliable, and secure sensor nodes. SET will sell integrated hardware boards and software IPs directly to companies that design and deploy sensor networks for applications like environmental monitoring and security.
1. Total number S treamlined
E mbedded
of customer
interviews
90
T echnologies
Integrated Hardware and Software Solutions for Complex
Sensor Networks
Click to view: SET Story Video
Chung-Ching Shen Shuvra Bhattacharyya Dean Chang
(EL) (PI) (CM)
Version: 5/23/2012
2. OUR TEAM
Chung-Ching Shen, Ph.D. (Entrepreneur Lead)
Research Scientist of Elec. & Comp. Engineering
7 years of experience in wireless sensor network research and system
development
6 years of experience in advanced embedded software development,
optimization, and testing
Shuvra Bhattacharyya, Ph.D. (Principal Investigator)
Professor of Elec. & Comp. Engineering
Written 6 books and authored over 200 papers
Experience in industry with high-performance design
Dean Chang, Ph.D. (Mentor)
Director of Mtech Venture Programs and Technology Advancement
Program
CTO & GM Gaming of venture-backed, Stanford-based startup Immersion
(NASDAQ: IMMR) in field of robotics/haptics
Version: 5/23/2012
3. OUR JOURNEY
3/20 I-Corps begins
3/22 1st pivot Sell embedded system design tools
Sell software IP for 4G
4/4 2nd pivot
broadband for cell phones
Sell software components , APIs, and
middleware for sensor networks
5/10 3rd pivot
Sell integrated hardware
boards and software IP for
select sensor networks
Version: 5/23/2012
4. BUSINESS MODEL CANVAS – VERSION 1
NSF I-Corps 2012 The SET Team
Technical support
and training* TS emphasis
Software tool Forums
development Platform/Chip makers
Platform providers Online community
Support / - TI/DSP, Nvidia/GPU,
- TI (multicore/DSP) DSP coders hate building
development iterationwrestling with clumsy Xilinx/FGPA, etc.
- Xilinix (FPGA)
- Altera (FPGA) tools that sidetrack
Tool providers
- NVIDIA (GPU) them from solving the
- Mathworks/Matlab, NI
core, tough problems.
/Labview, etc.
Tool providers Direct sales
- Mathworks SET
1D DSP developers
- NI People automates/optimizes OEMs - Telecom, SDR/CR, W
- Agilent Software IP tools for faster, cheaper (e.g., Mathworks) SN
DSP designs with
systems, speech, etc.
superior performance VARs (e.g., Xilinx
distributors)
License fees
People Subscription fees / per seat
Service fees
Version: 5/23/2012
5. SO HERE’S WHAT WE DID
Talkedto DSP engineers/researchers across
industry
Talked to Executives/Managers across industry
Version: 5/23/2012
6. SO HERE’S WHAT WE FOUND
Customers don’t want to learn new tools
“Big 3” have destroyed value chain
Version: 5/23/2012
7. OUR JOURNEY
3/20 I-Corps begins
3/22 1st pivot Sell embedded system design tools
Sell software IP for 4G
4/4 2nd pivot
broadband for cell phones
Sell software components , APIs, and
middleware for sensor networks
5/10 3rd pivot
Sell integrated hardware
boards and software IP for
select sensor networks
Version: 5/23/2012
8. SO HERE’S WHAT WE DID
Talked
to Chief Scientist level experts across 4G
broadband industry
Version: 5/23/2012
9. SO HERE’S WHAT WE FOUND
$25B (TAM)
Total embedded software engineering market
< $11B (SAM)
Market size looks great LTE embedded systems (LTE infrastructure - $11B)
~ $1B (Target)
(10% of SAM)
Source: MarketsandMarkets
But,
dominated by BIG players like Qualcomm
hard for startups to break in
Need $25M and “LTE guru”
Version: 5/23/2012
10. OUR JOURNEY
3/20 I-Corps begins
3/22 1st pivot Sell embedded system design tools
Sell software IP for 4G
4/4 2nd pivot
broadband for cell phones
Sell software components , APIs, and
middleware for sensor networks
5/10 3rd pivot
Sell integrated hardware
boards and software IP for
select sensor networks
Version: 5/23/2012
12. PROBLEMS WITH TODAY’S SENSOR NETWORKS?
Sensor nodes use too much power
Unsecure sensor node communications
Insufficient sensor node processing capability
Unreliable wireless communication restricts number
of sensor nodes in a network
Version: 5/23/2012
13. SET OPTIMIZES SENSOR NETWORKS
With SET, sensor nodes
Use less power
Support strong security protocols
Have increased processing capability
Support more reliable communication protocols, allowing
much larger sensor networks
High performance
region
Version: 5/23/2012
14. SO HERE’S WHAT WE DID
Talked to customers and partners in the sensor
networks industry
Building Surveillance
Security Environmental
management monitoring
Cisco,
American International ATR, Luna Innovations, Nokia,
Radio, Esensors, MESO, ObjectVideo,
BEI Technologies, Johnson Controls, North Dakota State U,
Texas Instruments
Impact Technologies, PARC, Sensys Networks,
Invocon, PNNL, Schlumberger, Sensor providers
Luna Innovations, Sigma Energy SolutionsTRX Systems
Physical Layer Systems,U. of Maryland
Raytheon, Automotive Embedded platform
SafeMobile
makers
Medical Texas Instruments ATR,
Lotus PSI, BEI Technologies,
Caterpillar, Esensors,
Medtronic,
U. of Michigan Physical Sciences
TSI
Version: 5/23/2012
15. SO HERE’S WHAT WE DID
Talked to customers and partners in the sensor
networks industry
Building Surveillance
Security Environmental
management monitoring
Cisco,
American International ATR, Luna Innovations, Nokia,
Radio, Esensors, MESO, ObjectVideo,
BEI Technologies, Johnson Controls, North Dakota State U,
Texas Instruments
Impact Technologies, PARC, Sensys Networks,
Invocon, PNNL, Schlumberger, Sensor providers
Luna Innovations, Sigma Energy SolutionsTRX Systems
Physical Layer Systems,U. of Maryland
Raytheon, Automotive Embedded platform
SafeMobile
makers
Medical Texas Instruments ATR,
Lotus PSI, BEI Technologies,
Caterpillar, Esensors,
Medtronic,
U. of Michigan Physical Sciences
TSI
Version: 5/23/2012
16. PRODUCT, VALUE PROPOSITION, AND CUSTOMERS
Middleware and
SET Product Reusable Software
Subsystems OEM HW components
More energy
Value proposition efficient, reliable, a
nd secure
Customer
segments
E.g., AIR, Raytheon, Eaton
Companies who design and deploy sensor networks
Version: 5/23/2012
17. MARKET
$12B (wireless) + $?B (wired) TAM (needs to be validated)
Total sensor network market (wireless and wired)
$12B SAM
Wireless sensor network market by 2013
$1~3B Target (needs to be validated)
Environmental monitoring, security,
and surveillance Projected
90 hundreds of
WSN nodes deployed
80 millions by 2013
70
60
(millions)
50
40
30
20
10
0
2005 2006 2007 2008 2009 2010
*Source: ON World
Version: 5/23/2012
18. REVENUE MODEL AND CUSTOMER ACQUISITION
SET Sensor Node Product
Middleware and
Reusable Software
OEM HW components Subsystems
($100 COGS)
SET’s price $400
Year 1: Year 2: Year 3: Year 4: Year 5:
(3 Customers) (10 Customers) (30 Customers) (100 Customers) (200 Customers)
1K nodes 5K nodes 30K nodes 200K nodes 400K nodes
$400K $2M $12M $80M $160M
First target
customers
Leverage our partners’
existing customers
Version: 5/23/2012
21. BUSINESS MODEL CANVAS – VERSION 6
NSF I-Corps 2012 The SET Team
Iteratively
apply, test, and
optimize SET tools
Sensor and processes More energy Companies who
device/network efficient, reliable, a Technical design and deploy
Iteratively develop nd secure sensor collaboration sensor networks for
companies and assess networks
node/network
Embedded solutions • Environmental
Seamless monitoring
platform makers
integration of
sensor interfaces • Security &
Application
and software Surveillance
specialists Embedded system subsystems
engineers Direct sales
System-level
software IP
People
Manufacture for large volume
Certification costs Selling integrated hardware boards and
Patents and sub-license costs software IPs
Version: 5/23/2012
22. BUSINESS MODEL CANVAS – VERSION 6
NSF I-Corps 2012 The SET Team
Iteratively
apply, test, and
optimize SET tools
Sensor and processes More energy Companies who
device/network efficient, reliable, a Technical design and deploy
Iteratively develop nd secure sensor collaboration sensor networks for
companies and assess networks
node/network
Embedded solutions • Environmental
Seamless monitoring
platform makers
integration of
sensor interfaces • Security &
Application
and software Surveillance
specialists Embedded system subsystems
engineers Direct sales
System-level
software IP
People
Manufacture for large volume
Certification costs Selling integrated hardware boards and
Patents and sub-license costs software IPs
Version: 5/23/2012
23. BUSINESS MODEL CANVAS – VERSION 6
NSF I-Corps 2012 The SET Team
Iteratively
apply, test, and
optimize SET tools
Sensor and processes More energy Companies who
device/network efficient, reliable, a Technical design and deploy
Iteratively develop nd secure sensor collaboration sensor networks for
companies and assess networks
node/network
Embedded solutions • Environmental
Seamless monitoring
platform makers
integration of
sensor interfaces • Security &
Application
and software Surveillance
specialists Embedded system subsystems
engineers Direct sales
System-level
software IP
People
Manufacture for large volume
Certification costs Selling integrated hardware boards and
Patents and sub-license costs software IPs
Version: 5/23/2012
24. BUSINESS MODEL CANVAS – VERSION 1
NSF I-Corps 2012 The SET Team
Technical support
and training* TS emphasis
Software tool Forums
development Platform/Chip makers
Platform providers Online community
Support / - TI/DSP, Nvidia/GPU,
- TI (multicore/DSP) DSP coders hate building
development iterationwrestling with clumsy Xilinx/FGPA, etc.
- Xilinix (FPGA)
- Altera (FPGA) tools that sidetrack
Tool providers
- NVIDIA (GPU) them from solving the
- Mathworks/Matlab, NI
core, tough problems.
/Labview, etc.
Tool providers Direct sales
- Mathworks SET
1D DSP developers
- NI People automates/optimizes OEMs - Telecom, SDR/CR, W
- Agilent Software IP tools for faster, cheaper (e.g., Mathworks) SN
DSP designs with
systems, speech, etc.
superior performance VARs (e.g., Xilinx
distributors)
License fees
People Subscription fees / per seat
Service fees
Version: 5/23/2012
25. BUSINESS MODEL CANVAS – VERSION 2
NSF I-Corps 2012 The SET Team
Integration with Technical
LTE LTE solutions for Tier I platform/chip
collaboration
Platform/Chip mobile phone and maker companies
makers SET development tablet makers that
have Tier II platform/chip
maker companies
higher
performance,
low power
LTE guru consumption, VARs (e.g. TI
distributors)
smaller
Software IP
footprint Direct sales
People
License fee / unit
Patents
Version: 5/23/2012
26. BUSINESS MODEL CANVAS – VERSION 3
NSF I-Corps 2012 The SET Team
Iteratively apply and
optimize SET tools Lower energy for
given functionality Technical
Iteratively develop
collaboration
and assess Higher functionality
Corporates (who
node/network for given energy
need sensor Corporate/environm
solutions
networks) ental networks
Small footprint
Sensor node
device providers
Embedded system
engineers VARs (e.g. TI
distributors)
System-level
software IP Direct sales
People
Manufacture for ?
large volume
Version: 5/23/2012
27. BUSINESS MODEL CANVAS – VERSION 4
NSF I-Corps 2012 The SET Team
Iteratively apply and
optimize SET tools Lower energy for Sensor network
given functionality Technical solution providers in
Iteratively develop the following markets
collaboration
and assess
Corporates (who Higher functionality
node/network • Building
need sensor for given energy
solutions automation
networks)
Optimization of the • Environmental
Sensor node
design trade-offs monitoring
device providers
Embedded system • Security &
engineers VARs (e.g. TI Surveillance
distributors)
System-level • Automotive
software IP Direct sales
People
Selling IP per unit/design/platform
Manufacture for
large volume
Version: 5/23/2012
28. BUSINESS MODEL CANVAS – VERSION 5
NSF I-Corps 2012 The SET Team
Iteratively apply and
optimize SET tools Lower energy for
given functionality Technical Sensor network
Iteratively develop solution providers in
collaboration
and assess the following markets
Sensor device Higher functionality
node/network
makers (e.g., high
solutions • Environmental
performance
Embedded processing) for monitoring
platform makers given energy
• Security &
Application Surveillance
Enhanced reliability
specialists Embedded system
engineers VARs (e.g. TI • Video networks
distributors)
System-level
software IP Direct sales
People
Selling IP per unit/design/platform
Manufacture for
large volume
Version: 5/23/2012
29. BUSINESS MODEL CANVAS – VERSION 6
NSF I-Corps 2012 The SET Team
Iteratively
apply, test, and
optimize SET tools
Sensor and processes More energy Companies who
device/network efficient, reliable, a Technical design and deploy
Iteratively develop nd secure sensor collaboration sensor networks for
companies and assess networks
node/network
Embedded solutions • Environmental
Seamless monitoring
platform makers
integration of
sensor interfaces • Security &
Application
and software Surveillance
specialists Embedded system subsystems
engineers Direct sales
System-level
software IP
People
Manufacture for large volume
Certification costs Selling integrated hardware boards and
Patents and sub-license costs software IPs
Version: 5/23/2012
30. VALUE PROPOSITION
SET Product
Middleware and
Reusable Software
Subsystems OEM HW components
Seamless integration
of sensor interfaces
and software
subsystems
E.g., ATR,
Esensors,
Physical Sciences
Companies who develop sensors
Version: 5/23/2012
Hinweis der Redaktion
When we first wanted to sell design tools, this was the business model. From this busy canvas, you can see that we knew this market very well. More specifically, we already knew many potential customers.
These are the customers we talked to.
We learned two things. First, our customers hesitate to learn new tools. But more importantly, our competitors, the three major companies, offered 95% discount to our potential customers and we simply cannot compete. This led us to make our first pivot.
So, to enter a more scalablemarket, we decided to sell software IP for 4G wireless communication.
We talked to manycustomers
and we found that the 4G market size indeed looks great, but the market is fully dominated and hard for startups to break in. Our customer told us to enter this market, we need $25 million and a guru
So, after another pivot, we explored the sensor network market. Originally, our plan was to sell software IP to this market. But after talking to the teaching team during the office hour and also discovered through customers that they are really more interested in software combined with hardware, we made our 3rd pivot.
In a sensor network, sensor such as cameras and microphones are distributed to monitor specific environmental conditions and communicate with each other.Examples: hurricane, at an entrance of an airport to recognize people’s face, on an UAV for battlefield, or around an oil pipeline
The current problems of sensor networks include power efficiency, security and reliability of the communication. Based on a customer we interviewed, the number of sensor nodes in a network is significantly limited to less than 30 because of these problems..
Our product can address these problems to enhance power efficiency and provide highly secure and reliable communication. As shown in our story video, our goal is to bring an existing sensor network solution to this so called high performance region so that the system resources can be fully utilized.With these enhancement, sensors can have longer battery life, and more sensor nodes can be used in a network.
Looked through 3000 conference papers in 2 days, found more than 100 authors from industry. Based on Steve’s suggestion, and it works pretty well
After processing carefully, we decided to select customer segments across the domains of environmental monitoring, security, and surveillance because we learned that the customers in these domains can benefit the most using our technology. We also found that sensor providers (based on the exceptional response rate) and embedded platform makers would be our key partners.
This is the summary of our product, value proposition, and select customer segments
The market is growing dramatically..
Progression: larger and larger sensor networks deployed