The organization

Stable, co-located teams

A team takes time to learn to work together, and it communicates better when its members are within earshot. Four patterns from A Scrum Book draw practical consequences from this: keep the same people together, place them in the same location, give visitors somewhere to wait and link teams through groups organized by trade. Remote work, which has become common since the book came out, means these patterns have to be reread with care.

Three projects and only one Sonia

Sonia is a data analyst at an insurance company. According to the allocation spreadsheet, she works 40% on the pricing project, 30% on fraud prevention and 30% on the management dashboard. She attends three team meetings a week, and allocations are reviewed every quarter.

A colleague from the pricing project drops by: "Could you look at my calculation this afternoon?" Sonia checks her calendar: "This afternoon I'm on fraud. Thursday, if that works for you." The calculation will wait two days. At the end of the quarter, none of the three teams can say what it will be able to deliver the following month, since its membership will have changed by then.

Stable Teams: keeping the same people together

A Scrum Book (2019) describes Scrum as a set of patterns: each one sets out a recurring problem, the forces at play and a solution drawn from experience. The Stable Teams pattern carries two stars, the rating for solutions the authors consider necessary in the long run.

The problem. Stakeholders appreciate teams that meet their forecasts, so the team tries to make those forecasts less uncertain.

The forces. Project management tends to treat people as "resources" to be allocated according to demand, with each new project or each crisis. This flexibility has costs: allocations have to be administered, every arrival forces the team to re-form and the newcomer has everything to learn. The book recalls Brooks's law, stated by Fred Brooks in The Mythical Man-Month (1975): "Adding manpower to a late software project makes it later." People shared between several teams, like Sonia, spend their time switching subjects. To regain some stability, organizations long froze requirements, which prevents learning along the way.

The solution. The organization keeps teams stable, avoids moving people from one to another and assigns each person to a single team whenever possible. It brings the work to the teams instead of rebuilding teams around the work.

The members of a stable team get to know one another and know what they can achieve together. Their velocity (the amount of work finished per Sprint, Scrum's fixed-length working period) then becomes meaningful, since it can only be measured if the same people stay together long enough. People who know they will keep the same colleagues are also more willing to invest in the way they work. The book cites a NASA finding reported in the Harvard Business Review in 2009: tired crews used to flying together made about half as many errors as rested crews who did not know one another.

Stability brings missing skills to light. The book advises temporarily assigning a specialist to the team that needs one, provided the specialist trains its members. The 2020 Scrum Guide does not deal explicitly with stability: it describes the Scrum Team, the whole team, as "a cohesive unit of professionals".

Collocated Team: within earshot

The Collocated Team pattern (two stars) complements the previous one.

The problem. Complex collective work demands very high-quality communication, and nobody can predict when it will be needed, how often or for how long.

The forces. Some people prefer to work alone, and a private office remains a sign of success. But as soon as people are separated, the small signs of colleagues' progress disappear, you have to book a room to talk and you fall back on poorer channels such as messaging or tickets. Yet communicating means being understood, which takes gestures, tone and an immediate reply. The book relies on Thomas Allen's curve (Managing the Flow of Technology, 1977), according to which the frequency of exchanges falls with distance: beyond about ten meters, the probability of talking at least once a week drops below 10%. A 2014 Harvard Business Review article argues that digital exchanges follow the same curve.

The solution. The team members are gathered within earshot, ideally in a room reserved for them. Alistair Cockburn gives as a benchmark that the whole team should fit within the length of a school bus. Quiet rooms remain available for phone calls or concentration. The book adds that if everyone spends the day with headphones on or communicates mostly by messaging, the benefit is lost, and that the Scrum Master (the person who helps the team practice Scrum well) should look for the reasons with the team.

The authors acknowledge that this way of working weighs on those who like solitude, and they go as far as writing that Scrum may not suit such people.

What about remote work?

The book came out in 2019, before remote work became widespread in 2020. Its position is clear-cut. Collocated Team asserts that physically separated members cannot work together effectively, and the Small Red Phone pattern adds that nothing makes up in the long run for development spread across several sites. An older pattern by James Coplien and Neil Harrison, Face-to-Face Before Working Remotely, asks at the very least that everyone be brought together in one place at the start of a distributed project.

The 2020 Scrum Guide says nothing about the workplace. It only states that "Optimally, all events are held at the same time and place to reduce complexity."

The evidence calls for nuance. Allen's curve describes a correlation observed in research laboratories and published in 1977. Since 2020 many teams have worked remotely or in hybrid mode, and the published results do not all point the same way. A study of more than 60,000 Microsoft employees in the United States (Yang and colleagues, Nature Human Behaviour, 2022) showed that the move of the whole company to remote work had made collaboration networks more siloed and exchanges less synchronous. It measures communication habits, not the quality of what is produced.

You can therefore take the pattern's forces rather than its single solution. The need for rich, unplanned exchanges remains when the team is dispersed, and it is up to the team to find how to meet it.

Oyatsu Jinja: a snack shrine for visitors

A team that works well attracts visits: other teams ask for advice, managers come to talk about their ideas. The Oyatsu Jinja pattern, or Snack Shrine, takes its name from the Japanese oyatsu ("snack") and jinja ("shrine"). It carries no star: the authors see it as one good solution among others.

The problem. The Developers need an environment free of interruptions, and yet they must exchange with the Product Owner (the person who orders what needs to be built), the stakeholders and other teams.

The forces. Isolating the team impoverishes communication, and opening its space to every visitor harms its effectiveness.

The solution. The team sets up, near its space, a corner stocked with snacks and drinks. It is far enough away that conversations do not disturb anyone, and close enough that people on either side can see each other. Visitors go there instead of walking up to the desks. If the person they are looking for is not there, they wait or come back, and the team members, who can see them, adjust the timing of their break. The place thus acts as the filter a person would otherwise provide. The authors liken it to a pattern by the architect Christopher Alexander, A Place to Wait (A Pattern Language, 1977), which asks that waiting be made pleasant.

Birds of a Feather: linking teams by trade

Stable, autonomous teams risk closing in on themselves. The Birds of a Feather pattern (one star, meaning a good solution that is not necessarily the only one) owes its name to the English proverb "birds of a feather flock together".

The problem. An organization's key skills cut across teams. Yet in teams that each bring together all the trades, the know-how of each trade only progresses through the team's own experience, and autonomous teams share little.

The forces. Empowering a team isolates it. Knowledge links form a network, whereas official structures are hierarchical. Everyone likes to progress in their trade, but one progresses little alone, and new ideas are born where paths cross.

The solution. The organization encourages informal groups for each important skill or area of interest, where the people concerned exchange on techniques, domain knowledge and methods. Participation is voluntary. A team that wants a say in a group agrees in return to follow its decisions. These groups manage themselves and have no authority over products: each team remains free to try a proposed improvement or not. Elsewhere they are known as guilds, quality circles or communities of practice, a term introduced by Jean Lave and Étienne Wenger in 1991.

In a hospital, for example, each department has its care team, and the link nurses for wound care meet once a month, across all departments, to compare their protocols. Each then goes back to her team, which decides what it adopts.

Limits and pitfalls

Stability does not forbid movement: people leave, others arrive, and the pattern only asks you to avoid needless moves. A team closed for too long can also settle into its habits, which trade groups and outside views help to prevent.

An open-plan floor of eighty desks is often mistaken for a co-located team. The pattern asks for a room for the team and quiet places beside it, which is almost the opposite.

The snack corner becomes a gimmick if urgent requests keep arriving through other channels. It deals with visits. Changes of priority go through the Product Owner, as the article Protecting the Sprint from interruptions explains. As for trade groups, they betray their principle as soon as they turn into committees that impose standards on teams.

Finally, the figures the book cites are second-hand, and how well they transfer to today's teams has not been established. The two stars reflect the authors' conviction. The article Criticism and the state of the evidence returns to this distinction.