Building High Performance Online Teams

What does a high performing team online look like? What measures should leaders adopt when monitoring and guiding their team’s performance? We’re about to take you on a deep dive into analysing what “good looks like” for high performing teams. 

This article builds on our introductory post on Building Better Online groups” where we identified that not all online groups are formed equal and that a more bespoke analytics approach is required for the different types of online groups. The ‘Team’ was one such grouping identified.   

Digging Deeper into Online Teams  

Until recent times, online teams relied on social networking, email or content sharing platforms for their online interactions. While these platforms provided a workable solution, the higher interaction frequency demands of teams provided an opportunity for new market entrants like SlackMicrosoft Teams and Hipchat to fill this need. Essentially, these team-focused platforms combine threaded discussions and real-time chat capabilities. Increasingly, they are now being positioned as ‘collaboration hubs’, from which team members can access other office work applications like calendars, online meetings, document stores and the like. 

As attractive as these emerging team platforms are, it is worth stepping back and asking the question; “What are we looking for from an online team?” Is it a new form of team that only now exists because the technology enables it? Or is it simply our current teams becoming more efficient by using a digital team-focused platform? What will success look like? How will we measure performance? How can we tell if a team is progressing well or not? Can we predict success from how a team is operating online? 

There is a plethora of research available on what factors makes a ‘non-digital’ team successful or not. It would be remiss of us to assume that because a team goes online the key attributes of successful off-line teams would change. Largely they do not. In a majority of cases, teams going online will enhance their effectiveness. However, there is also a risk that digitization may have a negative effect. For example, if your team members’ digital expertise is highly variable, it may result in the less capable team members being less engaged. Another area is team size. Digitization can result in larger teams, simply because it’s easier to invite more people in. A study conducted by IBM on its online groups found that 73 groups identifying themselves as a “team” had an average of 416 members! Extensive studies on team sizes found the optimal number was from four to six, with a maximum of 10 members. There is a good reason for this. We think of high performing teams having members who are connected in a reciprocal way to all other team members. In fact, the MIT study on the science of great teams identifies members connecting directly with one another – not just with the team, as a key attribute. In SWOOP terms, this would mean your %Two-Way connection for your team should be 100%. The larger your team, the more difficult this becomes. The graph below provides the simple math on the increasing effort required to sustain a cohesive team as team size grows: 

If we assume that to be a 100% engaged team, each member would have to have reciprocated interactions with every other team member. The total number of interaction is equal to n(n-1), where n is the team size. For a team of around seven members this is 42 interactions. For a team of 12 the number grows to 132, a team of 100 is 9,900 and for the average team size from the IBM study a whopping 172,640!   


We analysed more than 3,000 groups across 74 Yammer sites to show the following relationship between group size and group cohesion, measured as % Two-Way relationships. We can see the fall off in cohesion, with group size. This plot also shows that a small group size (more likely teams) does not guarantee cohesion, with many small groups having poor cohesion. The average group size was 624, with on average 137 over a six-month period. 

It’s not surprising that team cohesion drops with group size exponentially, so one must be very careful not to be seduced into growing a team as a larger group, just because digitization makes it easy.  

Jeff Bezos, the CEO of Amazon and the wealthiest person on the planet, was being prophetic when suggesting that a team was too large when it couldn’t be fed with two pizzas!  

Our first recommendation is to take heed of the research and limit your teams to two pizzas maximum. 

Multi-team Architectures 

In our analysis of 3,000+ active Yammer groups only five had 100% reciprocity; three had six members and the other two had eight members (and likely are teams). Of course, not all groups are formed as teams, but we suspect that a good proportion are formed as teams. If you are a leader of a 10+ member team it is now time to think about how you can be more effective by chunking your team into smaller, more effective teams, without compromising the overall intent and purpose of the larger group. Is there a core of members who are doing most of the work? Think about team outcomes. Can different teams be allocated to specific outcomes? What about those members, likely the majority, whom are simply invited in as passive stakeholders? Perhaps the existing large team could be declared a ‘stakeholder forum’, and then a number of smaller ‘doing’ teams created to work independently but aligned to the interests of the stakeholder forum. 

At the ‘doing’ end we are still likely to have multiple teams that need to be connected more intensely, especially if they are part of a value chain where one team’s inputs are dependent on the output of other teams; a so-called “team of teams” situation. We can learn something here from the software industry, where agile teams were pioneered. Large software projects were successfully delivered by multiple teams and hundreds, if not thousands, of team members. In this industry we have ‘connecting/bridging’ roles like ‘Program manager’, ‘Project manager’, ‘Solution Architect’, ‘Quality Assurance Manager’, ‘Technical Specialist’, whom can be found on multiple teams. Collectively, these team members form the bridges that connect teams, ensuring that their activities are appropriately synchronized.  


The formal ‘Project/Program Manager’ role would see these individuals being members of multiple teams, with a charter to ensure the interdependencies between teams are appropriately managed. Other specialist roles may also see individuals playing these roles on multiple teams, and hence in a position to broker shared practices between teams. Hence, the team of teams is created. The former large team still exists as a stakeholder forum; but now it is likely that there will be more tangible outcomes to review, discuss and provide feedback on. 

Team Level Analytics 

We have already spoken about the team cohesion measure of %Two Way Relationships, where we have identified the target at 100%. This measure will be sensitive to the time period that the team is being assessed over. The longer the period, the easier it is to achieve this goal. Therefore, it is important to decide on the appropriate team “heart beat”. High interaction teams like a quick response team may need to be assessed over a period of hours. On the other hand, a strategic planning team could be assessed on a weekly or monthly basis. 

Our second recommendation is to use the reciprocity measure of %TwoWay Relationships as the measure for cohesion in your team. Aim to get to 100%. 

Another team success factor identified by Google’s project Aristotle; a project formed to identify the common attributes of high performing teams at Google, is psychological safety. In essence, psychological safety exists when team members are able to surface concerns inside the team without fear of recriminations. High performing teams challenge each other by continually testing and questioning their performance. Research conducted on the team problem solving has identified cognitive diversity as a key predictor of success.  

Our third recommendation, therefore, is to select or encourage existing team members to participate in other activities or groups that can broaden their experiences and therefore their cognitive diversity 

The MIT study on the science of great teams suggests that team members periodically break, go exploring outside the team, and bring back information. 

Finally, we offer one final key performance indicator and that is the team’s curiosity i.e. to what degree do team members pose questions to the team? Unless the teams’ tasks are totally predictable, with each member’s role tightly programmed, one would expect a certain degree of questioning to happen within the team. Where a team is already cohesive, then the high levels of questioning may be largely related to agile co-ordination; a positive indicator. Where little cohesion exists a high level of questioning could be a negative factor, where team members may still not be aligned on purpose or job roles.  

Our final recommendation is to view your %Two-Way Relationship (reciprocity) together with your curiosity index (asking questions). Beware of a high curiosity score when paired with low reciprocity; where the team is likely non-productive, as it struggles to align its members. 

Recommendations for Building Better Online Teams  

To summarise, our key recommendations for developing productive online teams:

  1. Limit your online teams to the same size as your off-line teams. Research indicates that highly productive teams have less than 10 members, with a sweet spot of four to six members. 
  2. Aim to develop reciprocated relationships between all members of your team i.e. %Two-way connections for your team is 100%. 
  3. Select or develop diversity in your team members. Accept that to achieve diversity, team members must be given the flexibility to experience other contexts, either through participation in other teams or online groups.
  4. High performing online teams continue to challenge and question each other. Set a target of 15-20% of team online postings to be framed as questions. Beware that this target needs to be set in concert with the %Two-Way connections goal, to be effective. 

Subscribe To the SWOOP Newsletter

Join our mailing list, we regulary keep you up to date on the latest industry insights and research into Enteprise collaboration.

Thanks for Subscribing to SWOOP's newsletter