The Alibaba and AliExpress search bars are the doors to the thousands items sold by factories and resellers to a worldwide customer network. Algorithms control when, how often, and in which order products appear on the viewer’s screen, from clothing and furniture to tools and gadgets. In a way, the search algorithms on different e-commerce platforms are like environmental pressures that influence the survival and adaptation of the organisms (like objects) that inhabit them. Without access to the code itself, which is fiercely protected as intellectual property to deter competitors and prevent vendors from gaining an unfair advantage, the search algorithm can be studied only indirectly by testing its results.

Compared to Amazon or eBay, the search algorithm on Alibaba seems to put less priority on showing the most accurate results for the search query, thus freeing up space to display a variety of eye-catching and intriguing items to prolong visitor browsing. Yet these items are unusual in another sense: it is unclear if they are real products, potential prototypes, or simply illusions used to lure and entertain the online viewer. This collection of enigmatic products is materialised in paper, giving them a real (if flimsy) physical presence, while the metadata defining them as objects in the e-commerce platform is printed on the backs of the products.



Your work often deals with automation and machine-human interaction in design. Do you see the Alibaba sorting algorithm as a tool for design?

The sorting algorithm was more like our research subject or collaborator than a design tool: we are interested in showing humans how algorithms see objects. Commercial algorithms are fascinating, but they are very inaccessible to outsiders. The technology is protected as intellectual property to avoid being used by potential competitors. As designers, we could never work directly with the code, so we had to find other ways, like reading interviews and research papers to get insight into how e-commerce algorithms work. At the same time, the secretive nature of these algorithms gives them an interesting diversity between different platforms. Since Alibaba, Amazon, and eBay all write their own algorithms, they each have their own idiosyncrasies, like different species in the same genus, evolving in their own environments while also subject to the same global economic forces. For example, while Amazon’s algorithm seems to prioritise “accuracy” in terms of results that are most similar to your search query, on Alibaba and AliExpress the algorithm seems to want to captivate you with flashy and intriguing items to maximise the amount of time you spend on the platform. Is this a reflection of different attitudes towards consumption between China and western markets or just different technological strategies? There is no simple answer, but what is clear is that globalisation is still an abstract concept: digital tools remain very local and adaptive to their geographic habitats.

How does Alibaba’s algorithm change the conventional values, process, and relations associated with the modern design field?

Alibaba's algorithm presents the consumer with an impression of abundance and an extravagant kind of novelty. These “clickbait” objects make no reference to high-quality or detail-oriented manufacture in their attempt to lure us in. But when we went to China to investigate these objects in person, we discovered that many of them are custom-made or handmade. Alibaba’s platform takes the Toyota model of design and production to the extreme: many of these objects are only put into production once an order has been made online. The product image is more like a suggestion, and the vendor adapts it to the specific wishes of the customer. Alibaba facilitates this interaction by allowing consumers to chat directly with the manufacturer, even offering their own translation software. Of course, that makes Alibaba a significant part of the process: in a way, Alibaba is the designer guiding the interaction between the producer and the user through a technological system.

As human designers, how can you use, collaborate with, or hack the algorithm?

The strict definition of an algorithm is a set of instructions to solve a given problem, but given how they are implemented in complex technological networks, we should regard them as powerful actors in our society. The algorithm that determines which products appear on the front page of AliExpress is more like a car salesman than a library index. Its code executes a set of values—we may suspect we know some of them, but we cannot be aware of all of them. That is especially true with proprietary algorithms where we can only analyse the code through its indirect effects. The philosophy of open-source technology is empowering because it makes digital code available to the people it affects most—users as well as designers. But even if we could read the code, we would not be able to comprehend it fully. We need new tools to understand the complexity of such technology, which acts as a feedback system responding to the inputs of millions of users through forms of digital intelligence that are constantly being adjusted by developers. As more and more of our interactions and material acquisitions take place online, these algorithms will increasingly shape our understanding of objects and their intrinsic and social value. Our design process tries to start a dialogue with these digital actors and to give people a way to understand what they are and how they are affecting their lives.

Project Images

Copyright Design Academy Eindhoven. Photographs by Nicole Marnati.