Data Team Integration Models
Article written by Pete Crawford
Positioning data and analytics within the organisation.
Regardless of whether a data team is comprised of two, 20 or 100 people its ability to produce actionable insights and outcomes is heavily compromised if capabilities are not aligned to business needs. Furthermore, business needs are typically diverse, competitive and evolve as strategy unfolds and analytical use cases grow in complexity. To this end, it is worthwhile examining the differences between a handful of commonly used data capability models.
These models can also be viewed in the context of their suitability to divergent modes of business engagement, analytics maturity and resource coordination.
“Insights from Pete Crawford | Head of Data, Analytics & AI, Pete Crawford spends his day-to-day leading strategy, governance and execution over enterprise data platforms, data science and AI capabilities. Speaking at leading industry AI and data events, Pete is experienced in forming and directing multi-disciplined teams to manage enterprise information assets and deliver business outcomes through advanced analytics.”
Considerations
The adoption of a particular data and analytics organisation-alignment model is contingent on a number of considerations:
What data or analytics models can be shared?
What is the overall data governance structure?
What is the state of data infrastructure development?
How is data currently accessed and distributed across the organisation?
What is the analytics maturity or independence within other business functions?
Is innovation or AI a central tenet of business strategy?
Has there been consistent success with operationalising analytical models?
Common Options
The adoption of a particular data and analytics organisation-alignment model is contingent on a number of considerations:
What data or analytics models can be shared?
What is the overall data governance structure?
What is the state of data infrastructure development?
How is data currently accessed and distributed across the organisation?
What is the analytics maturity or independence within other business functions?
Is innovation or AI a central tenet of business strategy?
Has there been consistent success with operationalising analytical models?
Centralised
A single data and analytics team serves the whole organisation. Data people (analysts, data engineers, data science) sit together and treat other teams as clients.
Pros:
Sustainable funding; career growth aids talent retention; data is recognised as a strategic asset
Cons:
Lack of shared motivation or cooperation between groups; prioritisation challenges
Centre of Excellence
Combines the coordination of a centralised model with independent innovation intent.
Pros:
Focus and coordination when introducing new capabilities benefitting the rest of the organisation; optimal model for developing new infrastructure tools
Cons:
Can become isolated from business concerns; expertise may skew toward deep specialisations but neglect operationalisation skills; high operating costs.
Decentralised
Resources are funded and appointed by individual business units.
Pros:
Appropriate when there is limited inter-divisional coordination requirements (or inherent, irreconcilable internal conflicts)
Cons:
Duplication of resources; lack of ownership over data quality; data silos inhibit efficient data strategy
Federated / Embedded
Attempts to balance enterprise aspirations of the CoE model with the capacity to contract-out expertise for functional customisations. Analytics personnel report to business leads. Data engineering remains with the core data group.
Pros:
Encourages motivation and alignment of data-business objectives; retention of team identity; suitable for organisations with mature analytical competencies and systems
Cons:
Can lead to high costs; leadership conflicts between hub and spoke
Democratic / BI
Promotes self-service and domain-specific data ownership through development of data-as-a-service APIs and dashboards.
Pros:
Strong investment in data infrastructure; accessibility; rewards literacy with data visualisation
Cons:
High cost of infrastructure and training; systems need to be extremely robust as on-call engineering resources are scarce; limited role for data science and infusion of emerging data practices
Comparing Integration Models
Each model can be described and broadly evaluated against a set of coordination, management and capability factors.
A simple traffic light system denotes a generalised level of efficiency with deploying each model.
Mapping Integration Models
The five data and analytics integration models discussed can be loosely positioned in relation to both the complexity of an organisation’s analytical use cases and their capacity to control and coordinate data or personnel.
LOOKING TO Leverage and utilise your data? REACH OUT.
Whiteark is not your average consulting firm, we have first-hand experience in delivering transformation programs for private equity and other organisations with a focus on people just as much as financial outcomes.
We understand that execution is the hardest part, and so we roll our sleeves up and work with you to ensure we can deliver the required outcomes for the business. Our co-founders have a combined experience of over 50 years’ working as Executives in organisations delivering outcomes for shareholders. Reach out for a no obligation conversation on how we can help you.
Article by Pete Crawford
