As far as workplace trends go, several topics have dominated popular attention and discussion, from workspace design, digitalisation, Agile methodology, to the pandemic favourite of remote working. But there’s another important aspect of work that is being pushed increasingly to the forefront of organisation design: self-selection-based division of labour.

Essentially the process of matching workers with tasks, division of labour in organisations was, until recently, a top-down allocation exercise presided over by managers. These days, workers are increasingly empowered to assign tasks to themselves – a far cry from the pin factory labourers immortalised by Adam Smith. At software firm Valve and French auto parts maker FAVI, for example, employees select tasks as well as project teams based on their own perceptions of best fit. In fact, at many software firms and other less hierarchical organisations, self-selection has become the norm for its positive effect on employee motivation.

Yet traditional task allocation remains dominant and continues to flourish even in innovation-intensive sectors. Evidently, both methods of divvying up work have their merits and drawbacks. Knowing which to apply and when to apply it can have a direct impact on organisational performance and employee morale. 

In our new paper, co-authored with Massimo Warglien, we looked into the conditions in which self-selection would outperform the managerial allocation of tasks even if there were no motivational benefits: when employees are highly specialised, the tasks are independent of each other, and employee availability is unforeseeable.

As self-selection becomes more mainstream, our findings could help organisations aspiring to this process assess whether it is the right choice, or how it could be modified for a better fit.

Matchmaking employees with tasks

Existing research has established that, compared to traditional task allocation, self-selection of tasks benefits from employees’ own insight into their skills, which is often greater than managers’. Motivation, in turn, increases when employees are empowered to choose their preferred jobs. On the flip side, they often lack managers’ focus on organisation-level goals. Coordination may be compromised as a result. Some tasks may even be left unfilled.

To understand the implications of differences between the two approaches, we developed the idea of division of labour as a process of matching employees with tasks. We then built a mathematical model and used computer simulation to examine under what conditions either approach might enjoy a relative advantage.

We assumed that, in the managerial allocation process, allocators aim to choose employees from the available pool with the best skills for each available task. In the self-selection process, all employees pick the task for which they are best skilled.

When it’s simply better for bosses to allocate

It is obvious that managerial allocation outperforms self-selection in certain common division of labour scenarios. For example, when all tasks and employees are available and visible simultaneously, an allocator will aim to assign each task to the best employee and avoid understaffing, whereas self-selecting employees will pay little heed to the skills and choices of other employees. An analogous situation in the real world might be the launch of a new project with a given set of staff members.

When tasks come up at unforeseeable points in time, such as when an employee resigns, an allocator can select the best available worker. But if employees are allowed to volunteer for the job, a “pileup” may be the result.

Where self-selection may outshine allocation is when employees become available at different and unforeseeable moments. In situations like these, we found, the key factor that is likely to tip the scale in favour of self-selection is employees’ degree of specialisation.

The deciding factor for self-selection

When employees are highly specialised, our computations show, self-selection of tasks yield better organisational performance overall in terms of total skill level across tasks assigned. When employees are only moderately or not specialised, managerial allocation outperforms self-selection.

All other variables, including motivation and productivity, were kept constant in our model. Both the allocator and employees were assumed to be equally good at assessing skill for a task. Tasks were independent of one another and allocations were irreversible.

In practice, what our findings imply is that if you have many high-skilled employees with different specialisations, your firm will likely benefit from letting them choose the tasks they prefer. The overall quality of work and motivation will be high, and there will be little risk of a less qualified worker “blocking” a more qualified one from a task.

At the famed Dutch nursing organisation Buurtzorg, for example, self-managing teams of 12 nurses manage and conduct all tasks from providing at-home care to hiring, administration, scheduling and training; each nurse can choose and “craft” their portfolio of activities. On the other hand, if the employees have more similar skills, it might be better for a manager to allocate tasks.

Our study assumed that what employees are good at is also what they like to do. We could also extrapolate our findings to cover other variables:

• Managerial allocation for sociable workers – to avoid pileups by chatty types; self-selection for workers who prefer to work alone
• Managerial allocation when a talent pool is shallow; self-selection when a talent pool is deep
• Managerial allocation when tasks are interdependent, in which case any unassigned task could spell disaster; self-selection for standalone tasks.

Making self-selection work better

Division of labour through self-selection is an increasingly important topic for organisations, not least because younger people prefer to have more autonomy at work. But not all tasks or organisational set-ups lend themselves to self-selection. Our findings provide companies with a set of tools for determining whether this bottom-up approach would work for them.

Here’s the key question organisations ought to ask: Do our employees have highly specialised skills for the tasks on hand? Are the tasks more or less independent of one another? Is worker availability hard to forecast?  Proceed with self-selection if the answer to all of the above is “yes”.

Even better, modify the self-selection process with elements of conventional allocation. A hybrid process, if you will. Take Buurtzorg again. Procedures were explicitly designed to share unpopular administrative chores equitably among nurses. Another way to prevent tasks from going unstaffed is to encourage employees to pick those to which they could add the most value or make the most difference, rather than to choose those in which they are most skilled.

So instead of choosing Task A that Employee X has already picked (tasks can be performed multiple times in some contexts like software development), Employee Y takes unselected Task B where they can make a bigger contribution, even if Y is better than X at performing Task A.

Other hybrid ideas: encourage employees to take on “passion projects” of their own choosing; organise regular hackathons where anyone can pick and solve any problem from a list. Whatever form it takes, the division of labour remains as vital to organisation performance as ever.

Phanish Puranam is the Roland Berger Chaired Professor of Strategy and Organisation Design at INSEAD.

Marlo Raveendran is an Assistant Professor of Strategy at University of California, Riverside.