"An agile report compares agile scaling approaches in terms of their market penetration. In Section 6, NEXUS is mentioned at 1% first time this year, LeSS loses 1% to 3% and SAFe is the clear market leader with 28%. Between the frameworks, there are often bitter discussions in the appropriate forums on the Internet, who advocates the better scaling approach. This is in my opinion, but an unnecessary disgusting discussion, because
1. The differences in each implementation of a framework are strongly influenced by the corporate culture.
2. Frameworks are customizable by definition.
I therefore would like to deal with commonalities and the main emphasis of the three scaling methods.
It is e.g. an essential difference whether:
1. Three teams of seven developers working on one product - ,
or several 100 employees creating a solution with integration of hardware and software or the involvement of (agile) suppliers
2. The enterprise portfolio management should be agile.
3. The Scrum teams should be organized over a long-term budgeting along the Valuestreams.
4. The goal is to scale agile in a startup or in a young company. - or a classic enterprise like LEGO wants to be agile."
2. Kurt Jäger
Physicist…:
40 years of Software Development
15 years with agility & lean
10 years as consultant & coach
Management Consultant
kurt.jaeger@kegon.de
+49 611 20 50 80
Agile Management Coach e.g. at.:
Deutsche Bank Program Ltg. VK-Phönix
Deutsche Bahn Program Ltg. Step42
BMW.digital Agile Assessment
Audi & VW Customer journey
Software AG SAFe usage
CGM SAFe adaption
Generali Agile Transition
Societe Generale Agile Transition
KEGON AG
Biebricher Allee 119
65187 Wiesbaden
+49 611 20 50 80
www.kegon.de
KEGON AG 2016 2
5. Lieferung von Wert optimieren
Big Bang Große Inkremente Kleine Inkremente
Höchster Wert zuerst
Wert
Aufwand
Henrik Kniberg
Kosten
Value
Zeit
kumulativ
7. Warum agil/lean?
• Anpassungsfähigkeit an sich ändernde Welt - 3% Monat
• Komplexität nimmt zu
• Inkrementell - Fast Feedback
• Transparenz
• Wertorientiert priorisieren – das richtige tun – WSJF
• Waste reduzieren
• I&A
• Fertige Software als KPI
• Hohe Zufriedenheit durch Autonomie
• Hohe Produktivität
• Lösung 80% Scrum
12. Resource optimization vs Time-to-market optimization
Henrik Kniberg
C
Specialists
C D
TS
Cross-functional team
User needs
Specialized tasks
D
T
S
Resource optimization Time-to-market optimization
13. Feature teams
Cross-functional, self-organizing, colocated
Henrik Kniberg
Client team
C C C
Test team
T T T
DB team
D D D
Server team
S S S
Feature team 1
C
C
S
D
T
T
C
S
D
T
Feature team 2
D
S
DB
Server
Client
User
Communities
of interest
15. If Development Teams are not co-located:
Adjusted estimate = Estimate * 1.4
Adjust Estimates For Co-location
16. If more than 1 Scrum Team but fewer than 9:
Adjusted estimate = Estimate * 1.4
Adjust Estimates For Number Of Teams
Daily Scrum
Daily Scrum
Daily Scrum
Team 1
Team 2
Team 3
Coordinating
Scrum of Scrums
17. KEGON AG 2014
Warum Skalieren?
• Teamgröße ist limitiert 7 +/- 2
• Mehr als 1 Team arbeitet an einem Produkt
• 75% aller Projekte benötigen weniger als 3 Teams
• Siloteams – wegen Effizienzoptimierung aus der
Vergangenheit
• Feature-Teams ….
• Zu viele Manager
• Totslicen von Mitarbeitern
• Skalierng erst mal hinterfragen
18. Want to scale Agile?
DON‘T!
Craig Larman, author of Large Scale Scrum (LeSS)
Scale Agile?
19. KEGON AG 2014
Alternativen zur Skalierung
1. Silos auflösen mit inversem
Conway Law – Share Nothing Architektur
2. Große Projekte in kleine zerlegen
3. Gar keine Projekte mehr sondern jedes Feature einzeln
20. KEGON AG 2014 Scrum Day 2014 20
Conway‘s Law - rückwärts
24. Scrum Masterby Scrum.org – Improving the Profession of Software Development
Professional
Individuals and interactions over processes and tools
Working software over comprehensive documentation
Customer collaboration over contract negotiation
Responding to change over following a plan
25. Decouple as much as
possible
First the Spotify client was a monolith…
Feature Teams
Container teams
Now the client is a “container”
rik Kniberg
26. Release trains & Feature toggles
A
B
C
D
E G
F H
A B
C
E
C
E
F
G H
D
Week 12 Week 15
Release! Release!
26
Henrik Kniberg
27. KEGON AG 2014
Agile Methods and Practices
11th Annual State of Agile Survey 2016
28. KEGON AG 2014
Herausforderungen
agiler Skalierung
1. Agile und Lean Prinzipien: Value value Value…..
2. Aufbau-Organisation: Linie! HR BR
3. Selbstorganisation: … braucht Führung
4. Kultur: Agile Transition 1-5 Jahre
5. ASE: XP, DevOps, Agile Architecture
6. Kommunikation Reden, reden, reden, aufschreiben
7. Pull statt Push Runter mit der Auslastung
8. WIP Nicht alles gleichzeitig
9. Produkt-organisation Weg von Projekten
10. Tools: Atlassian, CA, versionone
29. KEGON AG 2014
Skalierungsstufen?
0. Scrum ein Team 1 Produkt - Maximale Autonomie
1. Agiles Produktmanagement - mehr als ein Team / Produkt
2. Agiles Lösungsmanagement – Boing 777
3. Agiles Portfoliomanagement – Discovery Kanban
41. Resource optimization vs Time-to-market optimization
Henrik Kniberg
C
Specialists
C D
TS
Cross-functional team
User needs
Specialized tasks
D
T
S
Resource optimization Time-to-market optimization
42. Feature teams
Cross-functional, self-organizing, colocated
Henrik Kniberg
Client team
C C C
Test team
T T T
DB team
D D D
Server team
S S S
Feature team 1
C
C
S
D
T
T
C
S
D
T
Feature team 2
D
S
DB
Server
Client
User
Communities
of interest
45. Squads are grouped into Tribes
Henrik Kniberg
Tribe Tribe Tribe
TribeTribe Tribe
Example: Spotify
46. PO PO PO
Tribe
Tribe lead
PO PO PO PO
Tribe
Chapter
Chapter
Tribe lead
PO
Chapter
Chapter
Guild
Each Tribe is a lightweight matrix
Vertical = Delivery.
Horizontal = knowledge sharing.
Example: Spotify