The power of algorithms & the case for String3

By Marshall Buxton

Why We’re Building String3

During the large-scale implementation of Historic Futures’ String2 technology, we signed-up and trained users in more than 700 different organisations forming part of the global supply chain of a major UK brand. Towards the end of the project, we noticed a significant trend. If we prepared information for a String2 user that compared their use of the tool (more accurately, the performance of the organisation they worked for) with that of a relevant peer group, they became noticeably more motivated to improve. Knowing someone else, somewhere in the world was doing “better” was a powerful motivator and seemed to change how our users related to the project. Our research suggests that this phenomenon is widely observed - and algorithms have become increasingly powerful drivers of behaviour as more and more relevant data is available to online systems. We’re working on an automated ranking algorithm for our new String3 service which will help all users understand how responsive they are compared to their peers.

Algorithms as Gatekeepers of Knowledge

In the broadest sense, algorithms are encoded procedures that transform input data into a desired output, based on specified calculations (Gillespie, 2012). Mathematical algorithms have been around for centuries, helping Babylonians decide points of law and Romans teach Latin grammar (Nunemacher, 2001). Today, these step-by-step procedures for calculations form the basis of computer programs and the logic that governs an increasingly computerised world. Algorithms help us navigate the digital world, recommend and curate content, and select what information is most relevant to the user. In this way, “algorithms acquire the sensibility of truth, because they ossify and calcify what is real” (Slavin, 2011). Said otherwise, algorithms are the gatekeepers to knowledge in the digital age.

Algorithmic Choice Architects

The power of algorithms to determine what information we’re presented with has far-reaching effects on our behaviour. This can be understood through the concept of choice architecture and the use of defaults. Choice architecture refers to the way in which environments, systems, and products are designed to direct people towards certain decisions (Thaler & Sunstein, 2008). Defaults can be defined as choice frames in which decisions are pre-selected, requiring no further searching or action from users. Defaults have powerful and pervasive effects on behaviour in that they’re hidden, non-conscious persuaders in most settings. For example, a building is built with doors, giving individuals a default choice of how to enter the building. In a similar way, algorithms display information to users in a certain format, directing their decisions by setting defaults and limiting the amount of choices available. A large body of research shows that consumer behaviour is easier to influence when less information is included, resulting in less choice. For instance, if Google’s algorithms provide less information to their users, they can direct them toward certain decisions and increase their revenue (Ghose, Ipeirotis, & Li, 2014). Degree Compass is another example of a product that explicitly uses algorithms and defaults in order to nudge users toward certain decisions and improved outcomes. Through an understanding of these two principles of behavioural economics, it is clear to see how inconspicuously persuasive algorithms can be.

Algorithms as Supply Chain Managers

In the context of supply chains, algorithms are used as a means of understanding complexity and making informed decisions. Algorithms are able to sort through and make sense of vast sets of supply-chain data, resolving uncertainty for the manager and directing them towards decisions. In the era of big data, where profit margins rely on exploiting increasingly large and complex data sources, algorithmic decision-making will only become more vital to supply chain management (Krishnan, 2013). Given the influence that algorithms have on behaviour, it is conceivable to imagine a not-so-distant future where algorithms have total control of supply chain decisions, making managers entirely reliant on their software, or replacing them entirely (Cohen, 2015).

Given the prevalence of algorithms in supply chains and their extensive ability to affect behaviour, we’re confident that String3 will nudge suppliers towards greater responsiveness.

String3 will be launched this Autumn. Sign up now to be the first to know when you can start using it

By Marshall Buxton @marshallbuxton & Tim Wilson