A recent Fair Work Commission decision has put algorithms front and centre in the contractor/employee battle, by finding that Deliveroo’s ability to ‘control’ a worker using its algorithm weighed heavily in favour of an employment relationship.
Last week the Fair Work Commission delivered its latest ruling on the question of whether a Gig Economy worker was an ’employee’ for the purposes of an unfair dismissal claim. The decision is significant in three ways:
- first, it is sure to raise eyebrows because – amid a relatively large number of unfair dismissal cases brought by Gig Economy workers – it’s only the second time the Commission has decided that the worker is an ’employee’ (and therefore entitled to the protection of unfair dismissal laws). The decision won’t be binding on future unfair dismissal cases, but it will be influential, and it shows just how thin the line can be between being a genuine ‘contractor’ or a deemed ’employee’;
- second, it involves close assessment about whether (in the Commission’s view) the worker was genuinely running his own ‘entrepreneurial’ business independently of Deliveroo; and
- third, it demonstrates that the courts and tribunals are developing a far more nuanced understanding of how company algorithms can be used to ‘control’ the behaviour of their workers (thereby indicating an ’employment’ relationship), and as algorithms are increasingly used to influence and direct worker behaviour, this will almost certainly have wider implications for the future categorisation of workers.
So – what happened?
Diego Franco worked for Deliveroo as a contract delivery driver from early 2017 until April 2020, when Deliveroo abruptly terminated his account for ‘failing to deliver orders in a reasonable time.’
Franco booked all of his delivery orders through the Deliveroo app which, during much of his engagement, used an algorithm that preferentially allocated work to delivery riders based on their individual performance metrics. These metrics included their ranked attendance rate, how many cancellations they had made at short notice, and their preparedness to work during peak times. A higher performance ‘rating’ resulted in (among other things) the ability to book work at preferred or optimal shift times when demand (and hence potential earnings) was typically higher.
Deliveroo also used data analytics software to log and compare rider delivery times, allocating a score to individual riders. Those whose low score deemed them to be too slow faced further investigation and, ultimately, termination of their engagement. It was this process that ultimately led Deliveroo to determine that Franco was slower than the other riders working in his zone and as a result to terminate his services.
Backed by the Transport Workers’ Union, Franco challenged his termination by filing a claim in the Fair Work Commission arguing that (despite ostensibly being an ‘independent contractor’) he was in fact Deliveroo’s ’employee’ and therefore protected by unfair dismissal laws.
The Commission’s finding – algorithms and the importance of ‘control’
When asked to determine whether a contractor is in fact a legal ’employee’, the courts (and in this case the Commission) assess the relationship against a range of factors, one of which is the degree of ‘control’ exercised over the worker. The more control that is exercised (or is capable of being exercised) the more likely that the worker is an employee.
Deliveroo argued that Franco wasn’t its employee because it didn’t control when or where he worked, or for how long he worked. It also argued that he could freely accept or reject delivery orders as he pleased, and (at least in principle) assign others to work in his place.
However, the Commission reasoned that while Franco may have appeared to operate relatively independently, the ‘practical reality’ of Deliveroo’s preferential allocation system meant that, to have any reasonable opportunity of earning an income, he was compelled to make himself available and to work at certain times, and not to cancel booked engagements. In this way, Deliveroo’s collection of individual performance metrics and other rider data (such as delivery times) in substance gave it an ‘extraordinarily vast repository’ of material that it could exploit to ‘exercise a significant level of control’ over its workers.
The Commission reasoned that this reality, combined with other considerations, meant that Franco was, on balance, Deliveroo’s ’employee’ and entitled to protection from unfair dismissal. This included the fact that Franco wore a Deliveroo uniform and had no capacity to negotiate the terms of his engagement, and could not possibly have developed any business ‘goodwill’ as a delivery rider independent of Deliveroo.
What does the decision mean?
The degree of ‘control’ over a worker will not on its own determine whether they are an ’employee’, but it does play a major part in that assessment.
What’s truly remarkable about this case is that, while previous decisions have taken into account the role that software platforms can play in influencing behaviour, few have considered algorithms in such an explicit and decisive way. The decision shows that courts and tribunals are developing a far more nuanced understanding of software platforms, and are likely to more readily reject the common claim by Gig Economy companies that they are merely software platforms that facilitate an exchange of services.
Further, whilst algorithms are at the heart of the Gig Economy model, the rise of ‘algorithmic management’ means that they are increasingly being applied in other workplaces. The decision therefore underscores the fundamental challenge many businesses are going to face in correctly categorising employees as ’employees’ or ‘independent contractors’.
Deliveroo has vowed to appeal the decision. This is perhaps unsurprising, given the implication that many of its other delivery riders may also be ’employees’ and hence qualify for minimum wage, leave entitlements and protection from unfair dismissal (among other things). But this isn’t without risk – as part of the appeal, it’s highly likely that the Commission will end up conducting a more forensic examination of Deliveroo’s algorithms and a further exploration of the degree to which they ‘control’ worker behaviour. We’ll be watching this space closely.
If you have any questions about your own operations, please feel free to contact our Employment and Safety team.
 Diego Franco v Deliveroo Australia Pty Ltd  FWC 2818.
 Diego Franco v Deliveroo Australia Pty Ltd  FWC 2818 at .