The general intention for Agile teams is that they are working in the same physical location, often called collocation, or being co-located. Much is to be said for the power of osmosis in terms of spreading knowledge among the team. Personal relationships are formed face to face, so that aspect is also strengthened by working in the same physical location. There is a lot more to it though. So let’s explore what science has to say about the real benefits of working in the same location?
The Agile Manifesto speaks about “individuals and interactions”, and the principals behind it weigh open and easy communication heavily as a factor for success.
Consistent with this, Osmotic Communication, as coined by Alistair Cockburn, refers to the free flow of information into the background hearing of members of a team — so that they pick up relevant information though osmosis. Team members seated in the same room may ask one another questions. The answers to which other members of the team may tune in to, contributing to discussion where the topic is relevant to their own work, as in the example below:
We had four people doing pair programming. The boss walked in and asked my partner a question. I started answering it, but gave the wrong name of a module. Nancy, programming with Neil, corrected me, without Neil ever noticing that she had spoken or that a question had been asked.
Cockburn contends that osmotic communication allows questions and answers to flow naturally and with surprisingly little disturbance among the team. He goes on to say that osmotic communication and frequent delivery facilitate such rich and rapid feedback, so that a project can operate with very little other structure.
He rated good levels of osmotic communication by the answer to two questions:
As you move away from this face-to-face situation, Cockburn believes teams experience a drop in communication effectiveness. As the richness of your communication channel cools, you lose physical proximity and the conscious and subconscious clues that such proximity provides. Cockburn’s predicted correlation between the richness of the communication channel and the effectiveness of that communication is demonstrated below.
The results of the Ambysoft 2008 Agile Principles and Practices survey which explored, among other things, the effectiveness of communication strategies between developers within an agile software development team and between team members and stakeholders, are summarized in the table below. Answers, rated on a range of -5 (very ineffective) to +5 (very effective), are fairly well aligned with Cockburn’s own predictions and make a pretty strong case for the necessity of colocation.
A common and proven approach for an agile team in helping to track and report its work is to use a physical task board. Usually done on a wall, the idea is to move task cards about between columns representing “to do”, “in progress” and “done”. Note cards are moved by the people who do the work and the resulting board represents the current state of all project work in an easy-to-interpret visual way. Teams practicing Scrum refer to this as the Sprint Board. This method is common across teams practicing Scrum as well as Kanban.
In a recent study by a small group of researchers in Finland, teams of masters students were instructed to use Lean Kanban in their software development projects. They then were subsequently interviewed about their perception of the benefits of this methodology. Researchers found there to be significant support for the success of the method:
Using a tool to replace a task board is generally seen to negate the benefits of the physical board. Teamwork software is altogether commonly thought to reduce information fidelity. Information can never be as effectively conveyed as in face-to-face communication. Teamwork software delays information transfer, and in some cases lead to information not being transferred at all.
Another research project — The Impact of Collocation on the Effectiveness of Agile is Development Teams — which used both quantitative and qualitative methodologies to provide insights into the optimal design of a collocation working space. It was carried out with two Cape Town based software development companies. Each employed the Scrum agile product development practices in a collocation setting. Data was collected through online questionnaires distributed to all co-located team members. These surveys consisted of statements to which respondents indicated their (dis)agreement using a five-point Likert Scale. Data was also collated via semi-structured interviews and observations.
Analyses of the data collected led to the conclusion that the Scrum software development teams working in the collocation environment were generally satisfied with the design of their co-located environment. Working in an open plan area and being co-located with other software development teams was seen to be beneficial. They confirmed that it led to increased knowledge and information sharing, and promoted the feeling of involvement and inclusion for each team.
A key recommendation arising from the findings is that managers should not only consider co-locating team members with each other. They should also investigate the possibility of colocation with other wider teams in an open-plan environment. Thus contributing to the feeling of connectivity and involvement that is shared within the environment.
A team practicing collocation means reduced coordination management. One common myth concerning distributed development is that the cost implications are no greater than the sum of the development expenses of each location. In reality, it is necessary to add to this the not insignificant overhead of collaboration and integration among the distributed team members. This includes additional labour costs, travel, as well as the introduction and maintenance of a set of tools. (Monica Yap, 2010) Adding to this we should also consider the cost of delay caused by the communication overhead. This results in value being released to the customer slower, thus reducing economic potential.
In this article I shared a small selection of the overwhelming evidence that shows how much more effective the colocation of a team is. It is my recommendation to seek collocation of a team wherever possible. Having the team together also helps a lot with the team’s growth in mastery. If it is not immediately possible I encourage you to work towards it.
In the meantime, or if you have no choice but to distribute your team members, I can recommend the body of work of Lisette Sutherland (no relations to Jeff Sutherland, in case you’re wondering). Lisette is a fellow member and supporter of Happy Melly where we help people be happy. She specialises in helping remote teams work better together.