04 fusionsmanagement (2007-04-26) - hasso plattner institute · © 2006, 2007 objectarchitects;...
TRANSCRIPT
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 0
Fusionsmanagement„Managing Application Portfolios in Merger
Situations“Vorlesung IT-Unternehmensarchitektur
VL 04; Donnerstag 03. Mai 2007; Raum HPI B-E.2
Fachgebiet Software-Architekturen, Prof. Dr. Robert HirschfeldDipl.-Inform. (univ.) Wolfgang Keller,
[email protected]://www.objectarchitects.biz/
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 1
Standort in der VorlesungImmer noch „Anwendungsportfolio-Management“
Stra
tegi
e un
d Pl
anun
gU
mse
tzun
gde
r Pla
nung
projectsProjects
Projektbegleitung (ARC6)
IT Strategie (ARC1)Unternehmensstrategie
ableiten
Monitoring des Projektport-folios (ARC5)
Mod
ellie
rung
und
Ric
htlin
ien
Entwicklung und Durchsetzung von Richtlinien (ARC4)
Modellierung (ARC3)
IT-Strategie
IT-Anwendungsportfolio-Management (ARC2 )
IT-Strategie Ist
Soll
Proj
ekte
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 2
Contents
• What are M&A activities and why would people want tomerge companies
• Forces and Roadmap for the high level pattern language
• The large scale patterns
• More patterns for the detail work
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 3
Frequent Scenario
• Two companies A and B perform a merger
• Both companies have similar lines of business. An importantcomponent of the merger is that both intend to benefit from so called„economies of scale“. Usually it is an important aspect to save on ITcosts
• The question here is: Ho do you deal with this.
• This task is a very common one for IT Enterprise Architects.
• There are even management patterns for this class of tasks
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 4
Merger Activities are increasingFrom the News (1) ...
http://www.gelon.de/news/Europaeischer_Bankensektor_auf_Konzentrationskurs_21056.html
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 5
Merger Activities are increasingFrom the News (2) ...
New York (Reuters) - Hoffnungen auf Fusionen in der Biotechnologiebranche haben dem Handel ander Wall Street am Donnerstag Auftrieb gegeben.
Die Anleger honorierten zudem die Entlassung von zwei Managern des US-Chemieriesen DowChemical, der zuletzt nach einem Zeitungsbericht als Übernahmekandidat galt. Die Spitzenkräftesollen Details über eine potenzielle Übernahme des Chemie-Riesen durch Finanzinvestorenausgeplaudert haben. Größere Gewinne wurden jedoch durch Inflationssorgen verhindert, die dieUS-Notenbank am Vorabend geschürt hatte.
Der Dow-Jones-Index der Standardwerte gewann im New Yorker Mittagshandel um 0,25 Prozentauf 12.515 Punkte. Der breiter gefasste S&P-500-Index legte 0,33 Prozent auf 1443 Zähler zu. DerTechnologie-Index Nasdaq notierte 0,62 Prozent höher bei 2474 Punkten.
"Die Liquidität am Markt ist noch immer hoch, und es wird weitere Übernahmen undUnternehmensbeteiligungen geben. Das ist ein positiver Faktor für den Aktienhandel", sagte PhilOrlando von Federated Investors.
http://de.today.reuters.com/news/newsArticle.aspx?type=marketsNews&storyID=2007-04-12T174111Z_01_HUM263658_RTRDEOC_0_BRSEN-NYSE-ZF.xml
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 6
Merger Activities are increasingFrom the News (3) ...
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 7
A few keywords from these articles ...
• In some (old) markets, M&As are the only way left if youwant to grow significantly. You need to “buy” marketshare.
• In (old) markets people often hope to improve their costposition by buying a competitor and then realizingeconomies of scale
• In some (new) markets that face a wave of consolidation,M&As are a must if you want to gain market share in orderto just survive
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 8
There are more reasons to buy a companyFrom Economic Theory you might know terms as
• Horizontal diversification• Vertical diversification• Forming a Monopoly or Trust
• which should be prevented by “anti trust agencies”• ...
Any of these cases raises the same question:How do I proceed with the IT functions
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 9
Contents
• What are M&A activities and why would people want tomerge companies
• Forces and Roadmap for the high level pattern language
• The large scale patterns
• More patterns for the detail work
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 10
Forces
• cost, synergies and economies of scale• socio cultural forces
• who‘s the boss?• NIH syndrome
• risk• functionality• nonfunctional requirements
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 11
On the Nature of Forces ...
costrisk
functionality
flexibility
simplicity
NIHsolution
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 12
Example from the Financial Industry:Why would you want to merge two financial institutions
5%+
2,5 – 3,5%+
1,5 – 2%
typical edp cost / premium ratios
Small insurance<~ 1 billion Euro
medium insurance~ 5 billion Euro
Large insurance~20 billion Euro
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 13
Why would you want to merge two FinancialInstitutions
• multi-channel approaches open multiple new channels tospend more money on the same customers
• Low costs allow you to offer the same products for a lowerprice
• There are significant economies of scale in the financialindustry ..
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 14
There are significant Economies of Scale in theFinancial Industry ..
~ 1200employees
in edp
~ 1200employees
in edp
largeGerman bank
+ „large“Austrian bank
~ 1200+xemployees
in edp
merged bank
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 15
Socio Cultural ForcesFrom the News ...
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 16
Socio-Cultural ForcesWho‘s the Boss?
• Bank A „buys“ Bank B in a „merger of equals“• Bank A‘s managers are slightly more equal than
Bank B‘s
• Which system will be chosen?• => depends
• Who will head the joined edp department• => make a guess
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 17
Socio-Cultural ForcesNIH syndrome
• Whose solution will a developer prefer?
• His own solution?• Another developers solution?
• That threatens to put him out of work• That will put him out of charge• That means that he has to learn a lot of new things ..
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 18
risk: which migration scenario has lower risk?
System
A
System
B
System
B
System
A
System
B
System
C
t
After
Before
Migrating data from A to B
( or vice versa ) Migrating data from A and B
to a new system C
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 19
functionality: compare two established systems
system Acovers ~ 70%
system Bcovers ~ 70%
100% possible functionality
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 20
80/20 principlea.k.a. good enough
100% cost
100% functionality
60%
80%
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 21
The last two slides imply ...
• often it does not really matter which of two establishedsystems you choose – at least from a standpoint offunctionality ...
• never try to merge a 100% solution from two 70%solutions using code level reengineering ...
• never try to build a system with 100% functionality – it‘stoo expensive
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 22
and finally ...
• there‘s also technology and nonfunctional requirementssuch as ...
• technical quality• maintainability• ...• ..• .• object-orientation
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 23
This talk is a bit like a pattern language
Steamroller
Cherry Picking
New Town
GovernanceStyles
Ladder of Integration Keep the Data
Toss the Code
Beauty Contest
You have a recurring problem: Mergerand a bunch of recurring solutions
plus a context of supporting patterns
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 24
How will you profit from this?
Steamroller
Cherry Picking
New Town
GovernanceStyles
Ladder of Integration
Keep the DataToss the Code Beauty Contest
You will be able todiscuss merger
situations using atleast the following
terms …
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 25
Contents
• What are M&A activities and why would people want tomerge companies
• Forces and Roadmap for the high level pattern language
• The large scale patterns
• More patterns for the detail work
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 26
One for the Start. In any ScenarioKEEP THE DATA – TOSS THE CODE
migrate from migrate to
DATA
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 27
Pattern: Data Migration
App A
App B
App C
App A*
App B*
App C*
Bank A Bank B
keep the data
toss the code ...
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 28
Kind of Anti-Pattern: Beauty Contest
• People will often perform a „beauty contest“ betweenlegacies
• only to find, that each covers 70%-80% of desiredmaximum functionality
• but each one another 70%-80%
Legacy A Legacy B
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 29
After the merger a series of workshops startsand people start arguing for two years!
App A
App B
App C
App A*
App B*
App C*
Bank A Bank B
which is better?
politics!
NIH syndrome 1 yeardecision process
100% functionality
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 30
The Options for the Beauty Contest
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 31
The Options You HaveNew Town - Rebuild the Portfolio
migrate from company A
migrate from company B
migrate to brand new portfolio forcompany A and B
Data
Data
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 32
The Options You HaveCherry Picking
migrate from company A
migrate from company B
migrate to mixed portfolio of bestapps from company A and B
Data
Data
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 33
The Options You HaveSteamroller
migrate from company A
migrate from company B
migrate data of B to portfolio ofcompany A. Drop B‘s apps
Data
done
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 34
Why you would favour the“Steamroller” combined with a “Data Migration”
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 35
pattern: one infrastructurepattern: one application wins
App A
App B
App C
App A*
App B*
App C*
Bank A Bank BBank AB
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 36
one might think it is a good idea to use the best from Bank A andBank B ...
App A
App B
App C
App A*
App B*
App C*
Bank A Bank B
App A
App B*
App C*
Bank AB
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 37
but .. each system uses infrastructure like workflow, dialogmanagers, printing ...
printing system
workflowsystem
dialogmanager
Bank A
App A
App B*
App C*
Bank AB
printing system*
workflowsystem*
dialogmanager*
Bank B
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 38
therefore ..
• try to avoid a mixed portfolio• if the two systems are well established and support the
business in a sufficient manner, functionality is not really aquestion – because it does not matter, which 70% youhave ..
system Acovers ~ 70% system B
covers ~ 70%
100% possible functionality
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 39
other downsides of the „mixed approach“ are
• you need code level migrations to kick out one set ofinfrastructure ...
printing system
workflowsystemdialog
manager
Bank A
App A
App B*
App C*
Bank AB
printing system*workflowsystem*dialog
manager*
Bank B
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 40
Pattern: Early DecisionDecide Fast and Hard ...
App A
App B
App C
App A*
App B*
App C*
Bank A Bank B
which is better?
politics!
NIH syndrome 1 yeardecision process
100% functionality
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 41
What really happens dependson a bunch of Forces
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 42
Force: Power and Politics
• Certain options only work in certain power constellations• try a steamroller in a merger of two parties with almost equal
powers
A B! A? B? A! B
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 43
Force: Speed and Time to Market
• The new company should be „back on the market“ with fullstrength in a short period of time after a merger
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 44
Force: Costs and Risk
• Building a complete new application portfolio from scratchis often close to impossible for financial reasons
• plus there is a substantial project risk• plus new systems often need years to match a legacy in terms of
performance and functionality• Cherry picking requires rebuilding interfaces and also
migrations• plus you need to adopt systems functionally to a set of common
base functions• Steam Roller is cheapest in most cases
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 45
Force: Capability to Perform Future Mergers
• The new company should be able to perform the nextmerger within a short period of time
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 46
The ladder of integrationshould be considered in most scenarios
ConsolidatingERP Implementation Concepts
concentrating data centers and infrastructure pays very fast!consolidating application portfolios remains a tedious task!
Concentrating Data CentersConcentrating IT Infrastructure Tasks
Concentrating ERP - Infrastructure
Harmonizingcore business applications
the ladder of integration:steps get highergains get smaller(relatively)
relative ROI
inve
stm
ent a
nd e
ffort
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 47
Contents
• What are M&A activities and why would people want tomerge companies
• Forces and Roadmap for the high level pattern language
• The large scale patterns
• More patterns for the detail work
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 48
What we have seen so far ...
Steamroller
Cherry Picking
New Town
GovernanceStyles
Ladder of Integration
Keep the DataToss the Code Beauty Contest
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 49
Alternative Roadmap ...More Patterns
Keep the Data – Toss
the CodeOne
Infrastructure
Early Decision
ClearVision
Application Architecture
Application Map
Archetype[AOC00]
The Bridge to the New
Town [Kel00]
Keeper of the Flame
[AOC00]
Stepping Stone
[Dew99]
should have
leadsto
is collection of
Is derived fromoften leads to
facilitates
may usemay use
facilitates
One Application
Wins
facilitates
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 50
Keeper of the FlameBasis Idea• Problem: How do you maintain the
conceptual integrity of a system,especially a long-lived one, in the face ofpotentially many change requirementsover time?
• Create a role whose function is to ensurethat all subsequently proposed changesare in accordance with, or at leastcontinuous with, the original vision of thearchitect(s).
• Full Paper: Allan O’Callaghan: Patternsfor Architectural Practice;http://www.coldewey.com/europlop2000/papers/ocallaghan.zip
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 51
Clear VisionBasic Idea
• Problem: You need to communicateideas about where an applicationlandscape is moving e.g. after amerger
• Solution: Give people a clear visionabout the ultimate goal of theirjourney
• Source e.g. Wolfgang Keller, “A FewPatterns for Managing LargeApplication Portfolios”, EuroPLoP,Irrsee 2001.
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 52
Application Mapswill be dealt with tomorrow by André Wittenburg
Quelle Bild: Credit Suisse, Dr. Heinrich Krause, Vortrag aufTagung IT-Architekturen, Deutsche Kongress, Frankfurt 2007
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 53
The Bridge to the New TownBasic Idea• Problem: How can you replace a part of
an existing monolithic system with a newcomponent or a complete newapplication portfolio.
• Solution: Install the new system part inparallel to the old system and connectthe two using a Bridge. Cut off any writeaccess to the old system’s equivalentpart (the one to be replaced) and redirectit to the new system.
• Source Wolfgang Keller: “The Bridge tothe New Town”http://www.coldewey.com/europlop2000/papers/keller.zip
Legacy
Part to be
replaced
New system
partBridge
other
legacy part
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 54
Stepping StoneBasic Idea
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 55
Summary following the Line of Patterns
Steamroller
Cherry Picking
New Town
GovernanceStyles
Ladder of Integration
Keep the DataToss the Code Beauty Contest
You have avocabulary for
managingapplication
portfolios in mergersituations
© 2006, 2007 objectarchitects; Wolfgang W. Keller - all rights reserved 56
Fragen?
und wenn Ihnen später noch Fragen einfallen ....Wolfgang KellerobjectarchitectsLiebigstr. 382166 [email protected]