How we moved from a vertical organisation structure to a service oriented one where our Ops teams were embedded in the value-driven development streams.
2. Purely traditional, vertical reporting
lines mean
There are skills not being utilised in some teams
Other teams are then wasting time learning these skills
We only do ad hoc value stream work; and then
Traditional teams are constantly being interjected by
value driven streams of work anyway
That reliance on developers to do Ops work is still too
great (AMI Armageddon)
There is duplication of work and overlapping reports
8. ValCro Service Stack
Platform Services
Data Services IT/Microsoft Services
Software Engineering (e.g. Content
Delivery)
Asset Management
Quality Services
Technical Problem Management
Customer Services and Incident
Management
IT Security
9. Example ValCro Team Responsibilities
Platform Services: Complete system deployment,
CMDB, System Performance Engineering (non-
application improvements and data analytics),
System security
Data Services: data modelling, database
technologies, database administration, BI
Asset Management: Content ingestion, asset
supplier relationships, curation, royalty reporting
10. ValCro Teams Mean
Better utilisation of people and skills
Increased efficiency
Increased opportunity for engineers and
managers to learn
Reduced dependencies and blocker
Greater focus on delivering things of real value
Better validation of goals and results
11. Remaining Questions
Do architecture report into Phil directly, maybe as R&D and then
feed into Software Engineering and Operations Engineering
Current Dev Leads should become known as Tech Leads
Current Tech Team should become known as Tech Management
Where does value streams come from?
Tech Mgmt: business project work (this is current Tech Team)
Tech Leads: internal project work (this is the current Dev Leads)
Problem Mgmt: unplanned work
Who reports to who?
I think reporting lines are traditional stack
And ValCro teams are dual lead (I have some ideas on this)
All weekly reports should probably be done on-line so people can
collaborate on them together
Editor's Notes
#5: Technical Operations (OS, middleware, technology management, Infrastructure, technical problem management)
Business Operations (reporting, IT, change management, partner cost management and QA)
#6: Operations Engineering name for two reasons: to make an infrastructure move more appealing and because Netflix have a department of the similar name and this always goes down well
#7: Who services who. Not totally linear and needs better visualisation