A new Deliveroo app algorithm, nicknamed ‘Frank’, has arrived with important implications including in relation to the UK Taylor Review.
The algorithm uses machine learning to continually evaluate the most efficient way to assign orders to riders. It does this by considering a range of factors including meal preparation time, time of day, vehicle and distance. As a result, the average delivery time of Deliveroo orders has fallen by 20%. Quicker deliveries, in theory, also means more deliveries, letting riders earn more. This also has benefit in helping achieve a key recommendation of the UK Taylor Review.
The UK Taylor Review is a recent government review of modern working practices and what makes up ‘good work’. One of the central recommendations is to ensure that the average hourly pay of riders equals 1.2 times the UK minimum wage. With the introduction of this new algorithm, riders can earn more per hour and so be more likely to earn at this level. This is all whilst retaining the flexibility that is fundamental to the gig economy.
For the UK government, ‘Frank’ is beneficial in that it doesn’t contribute to undermining state regulations like the minimum wage, something many claim is at the root of the abusive gig economy. This helps to bring the modern, disruptive gig economy more in line with current labour standards, allowing them both to thrive without either being undermined. There are, of course, other issues such as a lack of sick pay and equipment costs that the government needs to consider. Nevertheless, ‘Frank’ is going some way in improving the situation for gig economy riders and improving their chances of earning the minimum wage whilst retaining work flexibility.
From a Deliveroo perspective, the introduction of ‘Frank’ shows they are at least trying to improve the pay situation in light of the Taylor Review. This is a wise decision given the UK government’s ‘serious consideration’ of the findings. Whilst the company might be more focused on the 20% reduction in delivery time, good for keeping customers happy, improving pay prospects for riders should always be at the forefront. Furthermore, with Jinn ‘pausing’ operations and no mention of any recent UberEATS developments (other than with McDonald’s) Deliveroo appears to be leading the way on improving rider pay and safety.
However, it is possible that ‘Frank’ will lead to greater waiting periods for riders. If the number of orders remains the same, and riders are completing them quicker, they’ll also be waiting longer for the next one. From personal experience, on some nights there is no shortage of orders and quicker deliveries to move onto the next one will be welcomed. However, Saturday and lunchtimes seem particularly quiet, with waiting prolonged periods of time not being uncommon. If Deliveroo can use the quicker delivery time to encourage new customers, then orders should increase and waiting periods be unaffected. Otherwise, the benefits of quicker delivery in terms of pay might be less than expected.
Also, no definitive list of what characteristics ‘Frank’ utilises is given. I’d personally be interested to know what variables go into the algorithm to determine who is given which order. Whilst some variables are of little control, such as the type of meal, others might enable riders to improve their prospects of being selected for an order.
When riders are free, we’re told to go to a central meet-up area that is claimed the best location for being assigned orders. However, if there are ten of us waiting and an equal chance of being assigned an order, would it be better to be more spread out, possibly closer to some restaurants in a popular area? Whilst a detailed breakdown isn’t necessary, a better understanding of what influences ‘Frank’ might go some way in reducing waiting times for particular individuals. Even if it’s of little use to riders, it puts Deliveroo in a better light if they are open about how their platform works.
‘Frank’ clearly has important ramifications in a number of ways; it could offer greater opportunity for a minimum wage whilst retaining the flexibility of the gig economy, even if waiting periods might increase. Hats off to Deliveroo for trying to improve the overall experience of the gig economy, Jinn, UberEATS and others appear almost silent on this front.
Just as I was finishing this post, Deliveroo have revealed new tech that should improve the restaurant experience of their platform. Expect a post on this soon! For more information on Deliveroo’s use of data in its operations, see this short article.
The Delivery Gig 