Team performance is at the heart of building great products. There are countless theories, books and opinions around on what makes for good team performance, but we need to look beyond that and use models that are based on meaningful research.
Using data (especially from large studies) to inform how we think about and improve team performance will allow us to build upon the work of others. It also means we can move away from who yells the loudest or dogmatically references the latest management fad.
When it comes to models of team performance that are backed by large studies, we found these five models through a combination of our prior knowledge, recommendations and searching. An overview of each model is put forward along with a brief analysis of the model as it relates to the performance of product development teams.
At the end of this article is a discussion about which model to use.
#1 Google’s Team Effectiveness Model
Google studied 180 teams (115 of which were in product engineering) in order to understand what made some more effective and successful than others.
Google distinguishes between “work groups” and “teams” by saying a team is highly interdependent, and work groups have the least amount of interdependence.
Researchers used quantitative data like lines of code, bugs, customer satisfaction and more. But, realising these measures can be inherently flawed, they also captured subjective input from three different perspectives: executives, team leaders and team members.
Through this research, Google found that the factors that mattered, in order of importance, are:
- Psychological safety: Team members feel safe to take risks and be vulnerable with each other.
- Dependability: Team members get things done on time and meet the bar for excellence.
- Structure and clarity: Team members have clear roles, plans and goals
- Meaning: Work is personally important to team members.
- Impact: Team members think their work matters and creates change.
Things that were not significantly connected with team effectiveness: Colocation of teammates (sitting together in the same office), consensus-driven decision making, extroversion of team members, individual performance of team members, workload size, seniority, team size, tenure.
This research stands out as the most recent and relevant to product development teams, given the number of teams included in the survey and the comprehensiveness of the analysis.
What is possibly missing from the Google research is the nature of the team’s fit with the task at hand—however, you could say this comes under dependability. Someone can’t or won’t be dependable if they aren’t the right fit for the task.
Accelerate analysed 23,000 survey responses from 2,000 organisations—from startups to large enterprises—on the performance of software development teams with a particular focus around DevOps.
The factors they determined as key measures of team performance were:
- Lead time: The time it takes from a customer making a request to that request being satisfied.
- Deployment frequency: How often your team deploys code.
- Mean time to Restore: How long it takes to restore a software service after it has failed.
- Change Fail Percent: Changes that make it to production that fail.
The key issue with Accelerate is that it has a limited focus on team productivity outcomes, partly due to the study’s roots in DevOps. Although they relate these factors of performance to commercial outcomes through their survey, the fact that their measures of performance don’t include a measure of the outcome underscores this.
That being said, for some teams in larger organisations, it can be hard to measure direct outcomes. These proxy measures can provide concrete factors, where you can focus your efforts to improve team performance. It also focused on very technical aspects that could be thought of as lagging indicators of success for a team when compared to the other models in this article.
#3 Five Dynamics of Teamwork and Collaboration
This model was developed based on twenty years of research, where Frank LaFasto and Carl Larson asked approximately 600 teams and 6,000 team members to assess themselves and each other. The research covered a wide variety of industries, from airlines to sports, insurance and utilities.
LaFasto and Larson share the model in the book When Teams Work Best.
The model has five layers that increase the likelihood of team effectiveness:
- Team Member: Teams are made up of individuals who have the abilities and behaviours that matter.
- Team Relationships: The team must have the right culture and trust for giving and receiving feedback. Relationships should be simple and easy, not complicated and hard.
- Team Problem Solving: Teams need to raise and resolve the real issues.
- Team Leadership: Leadership that focuses on the goal, helps with collaboration and relationships, builds confidence, has some understanding of the work at hand, sets priorities and manages performance.
- Organisational Environment: An organisational environment that promotes clarity, confidence and commitment.
The above summary doesn’t do justice to the additional information and analysis provided in the book, which explores how to consider and strategically improve each layer.
This model breaks out the different elements of team performance at a high level. It is good to include the organisational environment the team exists in, as this often overlooked factor has a significant weight on performance.
The challenges with this model are that at a high level, the elements don’t seem to be entirely discrete or mutually exclusive—for example, Team Relationships and Team Problem Solving seem to overlap. The high level elements on their own aren’t as clear as some of the other models, they require digging into the subfactors and tactics suggested in the book. In day-to-day work if this model for team performance isn’t simple, it will be difficult to implement and will likely get forgotten.
#4 T7 Model of Team Effectiveness
Korn Ferry’s T7 Model of Team Effectiveness has been validated with 303 teams and 3,328 people in 50 different organisations. A review of the model found it to be one of the most comprehensive assessments of team effectiveness.
The T7 Model divides the factors of team effectiveness into internal (to the team) and external.
The internal factors are:
- Thrust: A common purpose about what needs to be accomplished or team goal(s).
- Trust: Belief in each other as teammates.
- Talent: The collective skills of the team members to get the job done.
- Teaming skills: Operating effectively and efficiently as a team.
- Task skills: Executing successfully or getting the job done.
The external factors are:
- Team-leader fit: The degree to which the team leader satisfies the needs of the team members.
- Team support from the organisation: The extent to which the leadership of the organisation enables the team to perform.
The separation of internal and external factors is an attractive aspect of this model, along with the meaningful amount of data behind it.
#5 Discipline of Teams
Jon Katzenbach and Douglas Smith developed a model for team performance based on data from more than 50 different teams, composed of hundreds of people, at 30 different companies.
The elements of this model of team performance are:
- Skills: Team members need specific skills to accomplish their work.
- Accountability: Team members have mutual and personal accountability.
- Commitment: There is a common approach, goal and purpose that is meaningful to the team.
These are focused on the outputs of Personal Growth, Collective Work Products and Performance/Results.
It’s a fairly comprehensive model that, importantly, keeps it simple with only 3 factors. In comparison to the other models, the two main areas where it is lacking are (a) that it isn’t as explicit about trust (although it is somewhat implied) and (b) that it doesn’t reference the environment that the team is operating in.
Excluded Models and Research
The following studies were excluded from this analysis:
- Hackman’s research and model for team productivity: Excluded due to the focus on intelligence analysts only and the lack of size (only 64 analytics teams).
- 5 Dysfunctions of a Team: Was excluded because it wasn’t based on reviewable research, although the book is insightful.
- The GRPI Model of Team Effectiveness: We were unable to find the original work authored by Rubin, Plovnick and Fry in 1977. From what was found, the model appears to have been developed in a health care setting based on the experience of the authors and other literature available at the time rather than a meaningful data set from a cross section of teams.
Each of the models covered above would help you think about, manage and improve team performance.
If you had to make a choice of the best one to use, then it’s likely between the Google Model or the T7 Model. Both are backed by a meaningful data set, and:
- The Google Model stands out because of the simple way it conveys the key factors you can affect in order to manage team performance, as well as ordering how important each is.
- The T7 Model stands out because it also has a simple way of conveying the factors that influence team performance, and includes factors from the external environment.
The other models don’t present actionable factors in the simple and usable way that these two models do.
It’s fairly even but the T7 Model takes into account external factors (team leader, and organisational support) where as Google doesn’t. The T7 Model also feels more comprehensive, with specifics published against each aspect to help you dig in, take action or evaluate how a team is going.
For a product development team, you can then combine this with a little more lagging measure of output with the Accelerate measures of team performance.
CEO & Founder
Scott has been involved in the launch and growth of 61+ products and has published over 120 articles and videos that have been viewed over 120,000 times. Terem’s product development and strategy arm, builds and takes clients tech products to market, while the joint venture arm focuses on building tech spinouts in partnership with market leaders.