At Agile India 2019, Doc Norton shared why the Tuckman team formation model doesn’t work and described new reteaming models that are more people-centric and applicable to current agile teams collaboration needs.
The Tuckman team development model, published in 1965, identifies five stages of team development, which are forming, storming, norming, performing and adjourning. According to Tuckman’s theory, these phases are sequential, of similar duration, and all necessary before the team can reach their highest levels of performance.
Based on Norton's observations, the Tuckman model doesn't work like predicted. The linear phases are not undertaken in set periods of time. Teams may never get out of storming, or they sometimes transition into norming to return and stagnate into storming for long periods of time. Norton recognizes that teams commonly stay stuck between storming and norming and may never go beyond their original performance levels.
A research study performed by Monterey Naval Postgraduate School in 2007 showed that only 2% of the teams surveyed go through all first four phases in a linear progressions. The stages are not entirely invalid, but they are not linear, most teams do not go through all of them, and storming is not a phase, but a continual process that happens frequently and in a healthier manner when teams are performing.
Norton believes that the Tuckman model had persevered until today because of the belief that long lived and stable teams are correlated to greater levels of specialization, therefore more accurate velocity predictability. Stable teams deliver better, therefore organizations, which considered people as project resources, sought to keep these resources on waterfall projects for the longest time, or shift them around based on companies priorities.
When team members are assigned to a single team and possess all the skills required to swarm at a sustainable pace, they deliver better. They face few context switching tasks, little to no external dependencies, and can focus on effectiveness. But people assigned to single teams created a new organizational challenge: team silos. Teams' learning happened within these silos, and impeded teams from cross-pollination. Agile teams created subcultures, some positive, some toxic. Teams were better at delivering, innovating and learning, but the organization wasn't learning.
Norton shared reteaming models and identified four criteria leading to better teams' collective satisfaction and performance: autonomy, connection, excellence and diversity. Some companies like Valve and Spotify understood the importance of reteaming and developed models to support teams' fluidity. Valve introduced the cabals, self-selecting and self-organizing teams largely temporary. Spotify leverages the model of squads, tribes, chapters and guilds to support more fluid teaming and learning patterns. Some reteaming patterns are:
- Socialization where a team makes a concerted effort to onboard new team members, and through conscious practice, they become good at absorbing new team members.
- Mitosis where a team grows until it is large enough to split. The newly formed team is already accustomed to working together.
- Volunteer Fire Department team model where teams volunteer members to create a temporary task team to solve a key issue and then re-join their team.
- Trading Places where teams temporarily swap team members between other teams to share learning across the organization.
Dynamic reteaming is becoming the norm for organizations to thrive, as documented by Heidi Helfand in various case studies.