Building a scalable infrastructure for games which serve hundreds of millions of people is tough. But with each new game requiring slightly different features building that scalable infrastructure multiple times becomes impossible. In this talk Steven will share how his team works with all of the different game teams to build a shared infrastructure that empowers the game developers without restricting their ability to innovate.
- Defining the central technology product
- Handling requests from multiple different stakeholders
- Managing the big risk - lack of adoption
- Keeping the platform lean and agile
2. Graphics Research
Group
Physics for Games Multi-core scripting
for console games
Mobile marketing
Automation
Early-Stage VC Mobile Games
B2B B2B B2B (B2B)2C
3. • Founded in 2003
• Candy Crush Saga released 2012
• IPO in 2014
• Joined Activision Blizzard 2016
• ~2000 employees currently
A brief history of King…
King has offices and
studios in Stockholm,
London, Barcelona,
Malmo, Berlin, San
Francisco, Chicago, New
York, Los Angeles and
Malta.
21. Partner Success
Engineering
Product Management
Inverse Conway!
“Organizations which design systems . . . are constrained to
produce designs which are copies of the communication
structures of these organizations” (Conway 1968)
Ads Services Data Infra
Developer
Platform
Game
Engine
26. Planning
Principles
1
2
3
4
5
Reduce complexity & fragmentation
Focus on products & features with high impact
potential & high repeatability
Focus on differentiated capabilities
Enable all games to be operable and data driven
Minimize cognitive load
28. Q1 Q2 Q3 Q4
Capacity
???
Prod
A
Prod
B
Prod
C
Prod
D
Prod
E
Prod
F
Cust
1
Cust
2
Cust
3
Obj
X
Obj
Y
Obj
Z
Active Investment Maintain
29. Source: CC BY-SA 2.5, https://commons.wikimedia.org/w/index.php?curid=985491
30. Q1 Q2 Q3 Q4
Capacity
Prod
A
Prod
B
Prod
C
Prod
D
Prod
E
Prod
F
Cust
1
Cust
2
Cust
3
Obj
X
Obj
Y
Obj
Z
Active Investment Maintain
Tech Debt X
31. Q1 Q2 Q3 Q4
Capacity
Prod
A
Prod
B
Prod
C
Prod
D
Prod
E
Prod
F
Obj
X
Obj
Y
Obj
Z
Active Investment Maintain
Tech Debt X
Maintenance
1
11 2
2
2 3
3
Cust
1
Cust
2
Cust
3
32. Q1 Q2 Q3 Q4
Capacity
Prod
A
Prod
B
Prod
C
Prod
D
Prod
E
Prod
F
Tech Debt
40%
Maintenance
1
1
1
2
2
2
3
3
Cust
1
Cust
2
Cust
3
Obj
X
Obj
Y
Obj
Z
X
Active Investment Maintain
33. Q1 Q2 Q3 Q4
Capacity
Prod
A
Prod
B
Prod
C
Prod
D
Prod
E
Prod
F
Tech Debt
40%
Maintenance
1
1
1
2
2
2
3
3
Cust
1
Cust
2
Cust
3
Obj
X
Obj
Y
Obj
Z
X
Product Backlogs
Active Investment Maintain
34. Q1 Q2 Q3 Q4
Capacity
Prod
A
Prod
B
Prod
C
Prod
D
Prod
E
Prod
F
Tech Debt
40%
Maintenance
1
1
1
2
2
2
3
3
Cust
1
Cust
2
Cust
3
Obj
X
Obj
Y
Obj
Z
X
Product Stream A
Product Stream B
Product Stream C
Active Investment Maintain
35. Q1 Q2 Q3 Q4
Capacity
Prod
A
Prod
B
Prod
C
Prod
D
Prod
E
Prod
F
Tech Debt
40%
Maintenance
1
1
1
2
2
2
3
3
Cust
1
Cust
2
Cust
3
Obj
X
Obj
Y
Obj
Z
X
Product Stream A
Product Stream B
Product Stream C
1
2
3
X
Z
1
1 1
X
X
Y
Y
Z
2
2
2 3
3
3
Y
Active Investment Maintain
41. Familiarity Bias
We have perfect visibility
of customer requirements
Reliance on old networks
Uncommunicated changes
Misalignment in org
Too many channels
{
42. 44
Alignment
King Leadership
strategic investments
& company initiatives
Game Teams
roadmaps, releases,
requirements, projections
Tech Leadership
tech debt, migrations,
new platforms, centralization
Product Areas
dependencies, resourcing
prioritization
Legal Team
compliance,
regional legislation
Platform Partners
capabilities, support,
legislation, requirements
48. Internal
Customers
Observations
1
2
3
4
5
No customer discovery → priorities by alignment
No sales renewal cycle to drive focus on customer
retention → “account” management
No Marketing funnel → emphasis on visibility, adoption
& evangelizing shared platform
Visible OKRs → Potential benefit to aligned
Competition is DIY → set centralisation bar high