In any field, the recipe for cultivating innovation isn’t complicated: reduce the cost of trying high risk ideas. Where Moore’s law and personal computers and parallel processing have done this for software, rapid prototyping is starting to do the same for hardware.The gap is closing, but too often it looks asymptotic—in many cases, Andreessenian wisdom still rings true: hardware is hard. In her PRODUCTIZED talk, Emily R. Batt from KAYAK explores the factors that drive the difference in economics between hardware and software development: capital, scale, and lifecycle. The talk also describes the best practices from both worlds of hardware and software. Access full talk here: https://vimeo.com/242063182

1. THE BEST OF
BOTH WORLDS
Combining innovation
strategies from hardware and
software development
E M I LY B AT T
Senior product manager
Cambridge, MA USA

2. TWO YEARS AGO, I LEFT A
JOB I LOVED IN
MECHANICAL ENGINEERING.

3. TWO YEARS AGO, I LEFT A
JOB I LOVED IN
MECHANICAL ENGINEERING.
I WANTED TO LEARN HOW
TO BUILD PRODUCTS FOR
THE CONNECTED
WORLD…

4. import java.io.InputStreamReader;
import java.net.URL;
import java.net.URLConnection;
import java.util.Scanner;
public class URLExpSimple {
public static void main(String[] args) {
try {
URL mySite = new
URL("http://www.cs.utexas.edu/~scottm");
URLConnection yc = mySite.openConnection();
Scanner in = new Scanner(new
InputStreamReader(yc.getInputStream()));
int count = 0;
while (in.hasNext()) {
System.out.println(in.next());
count++;
}
System.out.println("Number of tokens: " + count);
in.close();
} catch (Exception e) {
e.printStackTrace();
}
}
}

6. DEVELOPING CLIENTS’
CORE TECHNOLOGY
CONDUCTING R&D
BUILDING PROTOTYPES
IDENTIFYING TECHNICAL
STRATEGY
PREPARING FOR
PRODUCTION

8. ANALYZING DATA
RUNNING A/B TESTS
IMPLEMENTING
MACHINE LEARNING
IMPROVING UI/UX
DEVELOPING
COMMERCIAL STRATEGY

9. 19 500E M P L O Y E E S E M P L O Y E E S
PAPER CONSUMPTION
1 1 L B S
0 . 5 L B S
6
1 2 3 4 5 6
INBOX COUNT -100%
POWER TOOLS L
+100%
PING PONG JMONTHS

16. IMPACT
P (FAILURE)
COST
IDEAL
SOLUTION

17. IMPACT
P (FAILURE)
COST
IMPACT
P (FAILURE)
COST
IDEAL
SOLUTION
INCREMENTAL
SOLUTION

18. IMPACT
P (FAILURE)
COST
IMPACT
P (FAILURE)
COST
IMPACT
P (FAILURE)
COST
IDEAL
SOLUTION
INCREMENTAL
SOLUTION
POWERFUL
SOLUTION

19. IMPACT
P (FAILURE)
COST
POWERFUL
SOLUTION

20. IMPACT
P (FAILURE)
COST
POWERFUL
SOLUTION

21. IMPACT
P (FAILURE)
COST
POWERFUL
SOLUTION
IMPACT
P (FAILURE)
COST
ITERATION
Developed with Paul Sabin, Fikst

23. • What does it take to
build a product?
• What tools are
absolutely necessary?
• What are the barriers
to entry in the space?
• What sort of resources
are needed on-hand?
• What does the
development timeline
look like?
• Where are the
traditional pitfalls in
development?
• What happens once a
product is launched?
• What does it take to
distribute a product?
• How many users are
expected?
• What sort of growth
and loads are
anticipated?

24. S C A L E
L I F E C Y C L E
Ask coders what they need to start a
product: a laptop.
Ask mechEs: no two matching
answers.

25. H A R D W A R E S O F T W A R E

26. H A R D W A R E S O F T W A R E

27. H A R D W A R E S O F T W A R E

28. H A R D W A R E S O F T W A R E
R A P I D P R OTOT Y P I N G …

29. 1974
First
personal
computers
1980
First
computer to
sell 1 million
models
1985
10% market
penetration
1986
First SLA, 3D
Systems is
formed
1987
40% market
penetration
1989
Stratasys formed
2013
Stratasys purchases Makerbot
for $403 million,
3D Systems revenue hits $500
million.
Formlabs’ Form 1 ships.
1994
Stratasys IPO, $6
million
1977
48,000 PCs
shipped
2001
125 million
PCs shipped
2012
Formlabs raises
$3M on Kickstarter

30. A R D U I N O
L A S E R S A U R
F D M
F I E L D - R E A D Y
P R O T O T Y P I N G :
P h a r m a C h k
10% T H E C O S T
20% T H E F O R M FA C T O R
90% T H E A C C U R A C Y
Case study: rapid prototyping applied
to a diagnostic tool. Compared to
state of the art:

33. S C A L E
L I F E C Y C L E
Software: Minimal investment to develop,
version control mastery.
Hardware: A variety of tools necessitate
investment and space: upfront costs are
high. We see software companies move
into hardware after unlocking revenue
streams.
Watch: Makerspaces lowering hardware’s
barrier-to-entry. CAD programs
borrowing the best of version-control
from software.

34. C A P I T A L
L I F E C Y C L E
Capital and scale are the taxiing and
takeoff of product development;
capital facilitates production, scale
facilitates profit.

38. Identifying the limited resource can reduce friction in
decision-making.
Overhead is expected to be
low, so generally, the limited
resource is time:
Time to develop, time to
build, time to run, time to
load.
Because longer lead times
are par for the course, the
limited resource is usually
money:
Production costs, material
costs, distribution costs.

39. S C A L I N G U N D E R L O A D
Flight searches from
Florida IP addresses
before Hurricane Irma

40. Software: Building product can have low
overheard. Distributing product can have
low overhead. Margins can be high.
Many software products can scale from
one piece of hardware.
Hardware: Scaling is costly. Distribution
expenses accumulate.
Watch: Traditionally “shipped” software
moving to subscription models.
C A P I T A L
L I F E C Y C L E

41. Timelines—and the role of iteration—
diverge.
C A P I T A L
S C A L E

42. S O F T W A R E
H A R D W A R E
Brainstorming and
customer validation
Brainstorming and
customer validation
POC
Prototype
MVP development
MVP launch
Data collection and
iteration
Prototyping and
iteration
DFM First shots
and fixes
Parts!

43. 2 0 0 7

44. 2 0 0 7
2 0 1 7

45. Software: Short development cycles means
MVPs in user hands sooner, creating greater
risk of poor launch strategy. Software lives well
beyond its launch; data drives updates.
Hardware: Physical prototypes do a better job
of getting ahead of technical and product risk.
Forward-compatible design positions products
well for connectivity.
Watch: More “connected” hardware continually
blending the line between digital and physical.
Software security concerns pervading
hardware.
C A P I T A L
S C A L E

46. C A P I TA L , S C A L E , L I F E C Y C L E
drive the difference in economics between
hardware and software development.
I T E R AT I O N
can be front-loaded in physical product
development with the use of rapid prototyping
tools. This makes learning happen earlier.
Software benefits from data-driven iteration
when releases can be frequent.
V A L U E
is unlocked when the cost of trying probable
failures is reduced. This helps expose technical
risk and product risk sooner.