Product Line Economics:
Markets, Products, Reuse Potential --- how to benefit from PL and when not to do it..
These are the slides for a course at JKU Linz. It gives a short (<1 hour) overview of various aspects of the business aspects of Product Line Engineering (PLE) ranging from incentives to do PLE, factors that impact the decision all the way to when not to do PLE.
2. Ziele
Goals of the Lecture
• Understand the importance of a business perspectives on SE
– relation to company strategy
• Understand cost vs. benefits of PLE
• Scoping as a form of targeting
• When not to do PLE?
Note: This is just an overview of some topics / aspects,
we will not present detailed methods
1
3. Ziele
A few words about me.
• Studied computer science
• ~10 years Fraunhofer IESE
• Interest in business aspects
• Ph.D. in the area of product line scoping
• Professor software engineering
• GC, SC member/chair SPLC
• Rarely, doing some consulting
(mostly on PLE)
2
4. Ziele
Product Line Economics
Company success: costs vs. revenue
Good software engineering ≠ company success
cost reduction more products
↯
This is overly simplistic, but a starting point
3
5. Ziele
Goals of the Lecture
• Understand the importance of a business perspectives on SE
– relation to company strategy
• Understand cost vs. benefits of PLE
• Scoping as a form of targeting
• When not do it.
Note: This is just an overview of some topics / aspects, we
will not present detailed methods
4
6. Ziele
Cost reduction as a PLE benefit
# products
effort
traditional
development
approaches
Initial
Invest-
ment
(big-bang)
product line
development
Break-Even
incremental product
line adoption
• Product line engineering saves beyond the break-even point (often 2-5 (~3) products
→ Big Bang is best?
• Incremental approaches help to
– address risks by catering for agility, but are in total potentially less efficient
– avoid being late to the market ; missed revenue / customers
5
7. 6
• Initially created in 1996:
– develop IT-solution (inventory management system) for a large
warehouse;
all new development team, little experience on how such systems look
– expand to other warehouses of the same mother company
– start to create management systems for joint venture
– …
– become independent company for inventory management systems that
has a systematic product line
Example (incremental adoption)
maxxess Systemhaus, today: BISON Deutschland
9. 8
• Initially created in 1999:
– Company with a lot of experience on finance information systems (since
late eighties and very good contacts in banking)
– Decision to create a new product line for „web-based finance-information
systems“
• New small group with experience in web-technologies
• Stealth mode development (product line infrastructure) for 1.5 years
• Motivator: once on the market → be fast
• Enabler: a lot of expertise and contacts in the banking domain
Example (Bing Bang adoption)
market maker Software AG; today: vwd GmbH
11. 10
Where do PLE costs come from?
1. Adapt the
organisation: Corg
Product Line Infrastructure
2. Built the product line infra-
structure (core-asset base): Ccab
3. Build product-
specific parts:
Cunique
4. Re-use
common parts:
Creuse
Four
constituents
K. Pohl, G. Böckle, F. van der
Linden, Software Product Line
Engineering: Foundations,
Principles and Techniques.
Springer, 2005.
This can provide basis for ROI / DCF
calculation (see also simple model):
))
i
(p
reuse
C
)
n
1
i i
(p
unique
(C
cab
C
org
C
13. 12
Two (major) modes:
A. market-driven – standardized products, products aim at specific
„market segments“
B. customer-driven – one product per customer
C. combinations exist (standardized basis, but individual systems are
tailored to customers)
Understanding companies
Note:
• In customer-driven mode: revenue scales linearly with products
• In market-driven mode additional products lead to diminishing returns
(product line cannibalization)
14. 13
Marketed (Software) Product Line:
A set of products that are marketed together as sharing a common set of
concepts or features.
Note:
• the definition is independent of the engineering approach
Example: different WiSo-products
Engineered (Software) Product Line:
A set of products that are engineered together so as to share major parts of
their implementation.
Note:
• this definition is independent of the marketing of the product line!
• products may belong to different marketed product lines
Example: different navigation systems by a producer, some may be sold
under different brands
Engineering product lines
special case:
hierarchical PL
15. 14
Market Strategies [Por99+KB01]
• cost leadership – the company aims at providing the product at the
lowest possible cost
• differentiation – the products of the company differ by a certain
aspect (e.g., service, brand name, etc.) from the products of the
competition
• focusing – the company focuses on a specific niche (providing better
products / service there)
• faster – the company creates new products faster
Company strategies
PLE typically
positive
PLE may be positive
PLE may be positive
PLE is (initially),
later positive
16. 15
Why do companies use PLE?
• Decrease the development effort per product
• Decrease the time to market per product
• Keep time to close customer issues constant (i.e., not proportional to the
number of products)
• Achieve higher product quality
• Standardize user interfaces
• Address labor shortages (decrease development effort)
• Decrease the time for certification per product
• …
Expected benefits of PLE
Option Value
17. 16
The relation between effort / time to market (TTM) / revenue is messy..
• The relation between effort and time to market is not 1:1
– depends on overall project pressure (available stuff)
– churn in the market
– ..
• The relation between TTM and revenue is also not easy
Revenue dependencies
Entry Time
R
e
v
enue
s
mature market
R
evenues
innovative market
Entry Time
Δ revenue
18. Ziele
Scoping aims at
• identifying an optimal product set for the PL
• identifying optimal parts of the product line for reuse
17
19. 18
• Product Portfolio
– Which products?
– Which features?
– Product Portfolio Management (marketing science)
• Domains (conceptual units)
– What should be considered part of a domain?
– What`s in / what`s out?
– Domain Scoping (domain engineering)
• Assets
– What should be made reusable?
– Asset Scoping (product line engineering)
Types of scoping
They are all complex
in their right.
20. 19
• Typically strongly manual
• Only few methods
• some include quantitative elements (e.g., optimize in relation to costs)
• mostly qualitative (e.g., based on assessment methods)
Scoping methods
– Maturity
– Stability
– Organizational Constraints
– Resource Constraints
– Market Potential (Internal)
– Market Potential (External)
– Commonality / Variability
– Coupling and Cohesion
– Existing Assets
Viability
Dimensions
Benefit
Dimensions
Example factors
influencing
reuse potential
22. 21
• Being able to offer variability is good?
• We heard about diminishing returns of variation (PL cannibalization)
BUT: Sometimes variability is bad!
The benefits of PLE
23. 22
In the late 90s and early 2000 NOKIA was THE mobile phone
company
• widely admired for the usability of its phones
• extremely competent in PLE (hundreds of product versions per
year)
• Offered:
– app store
– touch control
– smart phone capabilities
in different phones!
Nokia
• Offered programming interfaces: no two phones offered the same set!
• Some companies made it a business to enable “applications” to work
on multiple phones
24. 23
• Phones at that time were thought of as products, not product
platforms / core of an ecosystem
• Developers were not regarded as important stakeholders in the
ecosystem!
Paradigm shift: the smartphone as a
compute platform!
• 1 model at a time
• developer friendly:
create once, run everywhere
• Developers HATE variability they need to take care off
The paradigm shift
In January 2007 the first iphone came out, Nokia had about 50%
of the market; in 2012 Nokia was below 3% world-wide
credits:
Matthew Yohe
en.wikipedia
25. 24
• Good platforms provide the basis for a software ecosystem
• Development partners in the ecosystem value if they can develop
once and run everywhere
• Users in an ecosystem value if they can learn s.th. once and can
always work with it
Android as an ecosystem
• Initially, a lot of variability possible (different phone makers), but
they are trying to improve on that.
Platforms as a core in an ecosystem
In ecosystems any explicit variability is disadvantageous!
If the business model of your ecosystems requires platform
variability, try to hide it or provide tools to cope with it
26. Ziele
Summary
Software Engineering (including PLE) is a large toolbox.
Understanding when each can be used beneficially
relative to a (business) context is as important as being
able to competently apply them.
25