Strategic Team Topologies for Large Agile Units
NielsenIQ agile coaches discuss managing 25-person teams in a market research context. They explore 'splitting without splitting' by using fluid backlogs, visualization of communication patterns, and context-specific frameworks to balance deep work with cross-functional cohesion.
The Challenge of Scaling Specialized Teams
In large enterprises with diverse technical roles, traditional agile scaling often fails due to rigid structures that ignore complex interdependencies. NielsenIQ, a market research firm with a significant technology operation, faced this challenge with teams of up to 25 members comprising data scientists, engineers, and QA specialists. The core tension lies in balancing the need for deep work in small groups with the necessity of cross-functional collaboration for a unified product.
Strategic Framework: Splitting Without Splitting
The agile coaches at NielsenIQ propose a nuanced approach: rather than permanently splitting large teams, they create fluid structures that allow for temporary separation. A key example is the use of three distinct backlogs (business operations, data science research, and engineering standards) within a single team. Members re-team every sprint based on priority, ensuring that resources align with immediate deadlines while maintaining a shared team identity. This method avoids the cultural fragmentation often associated with hard splits.
Visualization as a Decision Tool
A critical component of this strategy is the visualization of three forces: architecture, product, and people. By mapping communication patterns through one-on-one interviews and contrasting them with system architecture diagrams, teams can identify natural collaboration clusters. This process reveals whether a split is necessary or if internal 'deep work' zones are sufficient. The coaches emphasize that Conway's Law dictates that social structures impact technical outcomes, making human dynamics a first-class citizen in architectural decisions.
Metrics and Continuous Improvement
Success is measured not just by velocity but by developer experience (DX) and DORA metrics. These tools provide objective data to validate experiments, such as the adoption of lightweight LeSS elements to reduce ceremony overhead. One team saw a 63-point increase in satisfaction after splitting sprint planning into collaborative 'what' and autonomous 'how' phases. The approach underscores that there is no one-size-fits-all solution; instead, organizations must continuously experiment and adapt based on contextual data.
Conclusion
Effective scaling in complex environments requires moving beyond dogmatic framework adoption. By prioritizing visualization, fluid resource allocation, and context-specific experimentation, organizations can maintain both agility and cohesion. The key is to let the team drive the structural decisions, supported by data and facilitated by coaches who provide tools rather than mandates.
Key insights
-
Permanent team splits often destroy the cross-functional knowledge base required for complex systems. Fluid re-teaming based on sprint priorities allows for deep work without losing organizational cohesion.
Impact: Reduces knowledge silos and improves adaptability to shifting business priorities in large, multi-disciplinary teams.
-
Visualization of communication patterns and architectural dependencies is essential before restructuring. It reveals hidden dependencies and natural collaboration clusters that dictate effective team boundaries.
Impact: Prevents failed restructurings by aligning organizational design with actual workflow realities rather than theoretical models.
-
Framework adoption should be context-specific rather than holistic. Cherry-picking elements from SAFe or LeSS allows teams to address specific pain points without incurring unnecessary bureaucratic overhead.
Impact: Increases team satisfaction and efficiency by reducing ceremony fatigue while maintaining necessary coordination mechanisms.
-
Human dynamics are a strategic force equal to product and architecture. Ignoring interpersonal relationships and collaboration preferences leads to resistance and reduced performance during organizational changes.
Impact: Enhances trust and buy-in during transitions, ensuring that structural changes are sustainable and supported by the team.
-
Metrics like DX and DORA should be used as diagnostic tools for dialogue rather than punitive measures. They help identify root causes of performance issues, such as the trade-off between efficient processes and deep work.
Impact: Fosters a culture of continuous improvement and data-driven decision-making, enabling faster iteration on team practices.
Action items
-
Conduct one-on-one interviews to map individual communication patterns and collaboration frequencies. Create a visual graph of these connections to identify natural subgroups and potential bottlenecks.
Impact: Provides a factual basis for team restructuring decisions, ensuring that new structures align with existing social and technical realities.
-
Implement a multi-backlog system for large teams, categorizing work by domain (e.g., business, research, engineering). Allow team members to re-team each sprint based on backlog priority and required skills.
Impact: Optimizes resource allocation and ensures that critical deadlines are met without requiring permanent, rigid team splits.
-
Audit current agile ceremonies for efficiency. Split large events like sprint planning into collaborative 'what' sessions and autonomous 'how' sessions to reduce overhead and increase focus.
Impact: Reduces meeting fatigue and increases time available for deep work, directly impacting developer satisfaction and productivity.
-
Integrate DX and DORA metrics into regular retrospectives. Use the data to facilitate conversations about specific drivers, such as the impact of process changes on deep work capacity.
Impact: Creates a feedback loop for continuous improvement, allowing teams to validate experiments and adjust practices based on objective data.
-
Facilitate workshops that visualize the three forces: architecture, product, and people. Use these visualizations to guide the team in making decisions about team structure and collaboration models.
Impact: Empowers the team to take ownership of their working agreements, leading to higher adoption rates and more sustainable organizational changes.
Quotes
“So somewhat. Yeah. Maybe another point is like a lot of our data scientists also work in academics. So they are bringing in papers on their own work, they are sometimes professors.”
“So what did you find works best for you to uh uh uh deal with these data scientists and these multiple different roles? So maybe there is not one perfect solution. I think this is kind of the biggest realization that we've made that there is actually no perfect solutions.”
“So this is a little bit different thing to get your hand around is are we now one team? Are we two teams? And in the end, we are one team that has one team culture, one team name.”