In the agile world, there is frequent debate about whether shared services or cross-functional teams are the better choice. From the perspective of flow metrics and value stream optimization, there are clear scenarios where shared services provide advantages. This article explores when centralized expertise makes sense and when integrated teams should be prioritized— based on insights from flow metric analysis and evolutionary change management.

1. Shared Services: Cases for Centralized Expertise

1.1 High Specialization and Risk Management

Shared services are ideal when dealing with highly specialized skills such as compliance, security audits, or complex data analysis. The cost of duplicating this expertise in every team is often prohibitive. According to the Kanban Maturity Model, shared services enable risk mitigation by bundling and standardizing critical tasks.

Example: A central security team can analyze threats more efficiently and optimize incident response processes, while cross-functional teams focus on feature development.

1.2 Batch Processing and Scheduled Availability

When tasks require non-instantly available resources—such as database migrations or legal reviews—a shared service makes sense. The Kanban method recommends buffering and pull principles here to minimize wait times. For example: A shared legal team processes contract reviews in batches, rather than overloading individual teams with ad-hoc requests.

1.3 Economies of Scale and Utilization Optimization

With fluctuating demand or complementary workloads, shared services increase capacity utilization. Little’s Law (Cycle Time = WIP / Throughput) shows that centralized resources can increase total throughput, even with slightly higher transaction costs. A DevOps team providing infrastructure for multiple product teams is a classic example.

2. Cross-Functional Teams: When Integration Is Prioritized

2.1 Immediate Availability and Reduced

Coordination Costs

Cross-functional teams shine when fast feedback cycles and minimal wait times are critical. According to the Actionable Agile Metrics study, integrated skills shorten cycle time, as dependencies are eliminated. A Scrum team with its own UX designer and tester delivers features faster than if it had to wait for external services.

2.2 Cultural Factors and Trust

In high-trust environments, cross-functional teams enable autonomous decision-making. Kanban principles emphasize that teams with process ownership can respond better to customer needs. In conservative organizations with strict compliance requirements, however, shared services may remain necessary.

3. Flow Metrics as a Decision Basis

3.1 Work in Progress (WIP) and Cycle Time

An overloaded shared service is visible in the Cumulative Flow Diagram (CFD) through growing WIP bands. If the horizontal distance between “start” and “finish” increases, this indicates bottlenecks. Here, limiting WIP—a core Kanban principle—helps.

3.2 Throughput and Variability

Shared services should make their throughput transparent, e.g., via cycle time scatterplots. If variability is below 20%, predictability is high—ideal for plannable tasks. For innovative projects with high uncertainty, however, cross-functional teams are more flexible.

4. Practical Recommendations

1. Conduct Assessments

Measure the transaction costs for shared service requests.

◦ Analyze utilization of specialists vs. the need for immediate availability.

2. Test Hybrid Models

◦ Use shared services for compliance/infrastructure, but integrate frequently needed skills (e.g., testing) into teams.

3. Adjust WIP Limits

◦ Limit requests to shared services to avoid overload (e.g., Kanban WIP limits per service channel).

4. Establish Feedback Loops

◦ Hold retrospectives with shared service and product teams to identify bottlenecks early.

Conclusion

The decision between shared services and cross-functional teams depends on the degree of specialization, demand structure, and organizational maturity. Flow metrics such as WIP, cycle time, and throughput provide a data-driven foundation. Ultimately: Flexibility arises from clear prioritization and continuous adaptation—not from dogmatic models.

Actionable Agile Metrics for Predictability (Vacanti, 2016)

Kanban – Evolutionary Change Management (Anderson, 2010)

‍

‍