Individual Work - Runpei HE

This post will analyze data and algorithms used by different actors in food delivery, and discuss how they shape and are shaped by digital culture. This section begins with analyzing the main article “Exploring the inbound and outbound strategies enabled by user generated big data: Evidence from leading smartphone applications”, then evaluating it with a digital culture perspective, and linking it to other sources.

I selected this article because it uses a unique theoretical framework the provided a new perspective to appreciate the power of data. Furthermore, the author’s methodology combines rich literary analysis, qualitative methods and quantitative analysis to provide convincing results. The approach of comparing four data-generating applications also broadened my perspective.

This section will organize the relevant content of the article deepen our understanding of this blog’s topic.

In this article, various concepts are used to analyze food delivery applications and other data-generating applications. They can be applied to a wide range of topics on digital culture, so here is a summary of these valuable lenses of digital analysis:

• An Open Innovation Paradigm (OI) allows the more structured and systematic use of resources for innovation, and it is enabled by the Big Data collected. Here, the data is not limited to one company (restaurant), and the knowledge flows across companies. External source of knowledge – “users, suppliers and competitors” – all provide valuable insight. Later, these innovations can be shared by the company with the market to suit their diffusion.
• User Generated Big Data (UGBD) is information produced by users when they utilize and interact with existing digital products (e.g. food delivery apps). This is central to the innovation process, and can be analyzed using the Open Innovation Paradigm.
• The article repeatedly mentioned “innovation”, which is central to our understanding of changes in digital culture: innovation leads to new tools and usages, which then contributes to the evolution of culture and social practices in our daily lives.
• Human-centered design (also known as user-centered innovation) emerged in the late 1990s, aiming to put the consumer into context. In other words, the social and cultural context of the user determines how they will receive and use a product. The producers will consider the symbolic value of their products, and how it impacts the users. In this case, users (1) provide information and (2) benefits from the company’s products.

This article compared four applications providing User Generated Big Data. This blog will focus on Deliveroo, which is a food delivery start-up private company founded in London in 2013.

Deliveroo works with various sources of data: the customers, the restaurants, and the riders, but also external data on the time, geo-location, routes, etc. Here, just like human-centered design suggests, these three main parties – customers, restaurants and riders – both produce data, and benefit from the data produced.

Restaurants is the main focus in this article. This is valuabvle because previous posts in our blog did not focus on the importance of the app to restaurant, and the power of data generated by restaurants. This article reveals that restaurants benefit by receiving accurate time predictions, which is central for restaurants to know when the food needs to be ready, and when to produce it to preserve the best taste of food. Deliveroo’s algorithm also provides various advices to improve the business of restaurants, such as how to change the dishes, how to package, trends for specific dishes overtime, how a special dish can increase the flow of customers, how to select places to open new restaurants and expectations for potential requests. On the other hand, restaurants generate valuable data on the types of orders received, amount of orders, and their location distribution. These data helps other restaurants in aforementioned ways, while guiding riders to order-dense regions, and influencing the behavior of consumers when they decide based on the options made available by restaurants.

Comparted to other applications where users generate data – Twitter, Spotify and Strava – the authors concluded in the article that a few commonalities can be observed. Firstly, the users are not usually aware that they are contributing to the data production. Secondly, the context influences how various users utilize the applications. The authors concluded that User-Generated Big Data used for User Centered Innovation is central to these applications, replacing previous ways that data is collected by companies. In these examples, data directly creates values for the users of the applications. 

It can be observed that the data produced by the main parties – consumers, restaurants, riders and the platform – provides guidance for other parties to optimize their behavior. The experience with food delivery algorithms does not only actors in isolation, but contributes to a multilateral process. Here, an algorithm is absolutely necessary as the complex reality requires computer calculations to meet the requirements for accuracy and consistency.

I conducted additional research to deepen the analysis on Deliveroo and its use of data in its algorithms. Firstly, an analysis conducted by Ipshita Sen reveals that, in addition to the aforementioned uses of data to support decisions and recommendations, it conducts “real-time operational monitoring”. In my opinion, compared to previous usages of algorithms which emphasize future prediction models or historical pattern analysis, this real time response is especially valuable to respond to immediate demands of consumers, and to provide immediate information to guide riders and restaurants.

This reliance on real-time information contributes to shaping the “culture” and behavior in the food industry. For example, while consumers may previously rely on printed guides or written articles for restaurant recommendations, Deliveroo now has an additional function to recommend restaurants real-time in London. While restaurants used to rely on their staffs to have people decide on how to manage consumer preferences, demand and raw food stockage, now they rely on an application to adapt to the changing world. Riders, unlike traditional drivers, no longer need gradually finessed experience with the city to do complete their delivery. The application greatly reduced the training time required. In other words, the daily experience of people is shaped by food delivery services and the algorithm that stands behind it.

The algorithm also causes ethical concerns. For example, similar delivery apps in China makes accurate arrival time predictions based on rider performance, but this also means that if a rider was fast with food delivery, the algorithm would raise expectations, and penalize the rider if they fall below their previous performances. This creates ethical challenges, as the mental health of riders could be compromised on high-stress environments, and lead to increasing accidents. In Richardson’s article, she discussed how riders have little flexibility. The application requires riders to have continuous and long attendance in order to get priority for order offers, which contributes to the long and stressful hours of work. The algorithm may also provide false expectations to riders about which areas have higher demand, as discussed in other blogs.  

On the other hand, in my opinion, the algorithm can be applied in other immensely beneficial ways. By optimizing rider and raw material delivery routes, carbon emissions can be reduced. By better expecting consumer needs, perhaps restaurants can have more accurate calculations to guide their food preparations, and reduce food waste. Furthermore, research conducted by Nair et al. shows how this algorithm can be applied to emergency reliefs, such as delivering supplies to regions that just suffered from natural disasters.

Comparing different delivery strategies: 

Forecasting methods: 

Below are some personal reflections for this individual work blog post.

• This article is selected because it is needed to provide the basics for algorithms, which will be elaborated on in later posts.
• This post incorporates stylized texts, visual banners and videos to keep the content concise, relevant and engaging.
• Personal analysis and links to other academic works are included to make the analysis more meaningful. 
• See the Algorithm section for details on the impact of the algorithm on riders' decision-making. 

Daniel Trabucchi, Tommaso Buganza, Claudio Dell'Era, Elena Pellizzon. 2018. Exploring the inbound and outbound strategies enabled by user generated big data: Evidence from leading smartphone applications. John Wiley & Sons Ltd. Creativity and Innovation Magazine 2018;27:42–55.

Lizzie Richardson. 2020. Platforms, Markets, and Contingent Calculation: The Flexible Arrangement of the Delivered Meal. Antipode Vol. 52 No. 3 2020 ISSN 0066-4812

Manas Joshi , Arshdeep Singh, Sayan Ranu, Amitabha Bagchi, Priyank Karia, Puneet Kala. 2020. Batching and Matching for Food Delivery in Dynamic Road Networks

Newstex. 2016. Tech City News: How Deliveroo and UberEATS are changing the £9bn food delivery landscape. Newstex Trade &Industry Blogs; Chatham

Statistics provided by Statista

Business wire. 2019. Online Food Delivery Companies are Betting Big on AI and Machine Learning. Business Wire on Marketing Technology Insights. [Online] Available at: https://martechseries.com/analytics/online-food-delivery-companies-betting-big-ai-machine-learning/

Ipshita Sen. 2020.  How Deliveroo uses machine learning to power food delivery. Outside Insight. [Online] Available at: https://outsideinsight.com/insights/how-deliveroo-uses-machine-learning-to-power-food-delivery/

D.J. Nair a, H. Grzybowska a , Y. Fu b , V.V. Dixit. 2016. Scheduling and routing models for food rescue and delivery operations. Socio-Economic Planning Sciences