Profiling of users in online platforms

This signal post drills into the topic of profiling of platform users. We will have a look at how information on users’ background together with data on their offline and online behaviour is used by platforms and allied businesses. On the one hand, profiling allows service providers to answer user needs and to tailor personalised content. On the other hand, being constantly surveyed and analysed can become too much, especially when exhaustive profiling efforts across platforms begin to limit or even control individuals based on evaluations, groupings and ratings. The ever increasing use of smart phones and apps, as well as use of artificial intelligence and other enabling technologies, are in particular accelerating the business around profiling, and individuals as well as regulators may find themselves somewhat puzzled in this game.

What is profiling all about?

In the context of the platform economy, we understand profiling as collecting, analysis and application of user data as a part of the functioning of a platform. This means that e.g. algorithms are used to access and process vast amounts of data, such as personal background information and records of online behaviour. The level of digital profiling can vary greatly, and a simple example would contain a user profile created by an individual and their record of activities within one platform. A more complex case could be a multi-platform user ID that not only records all of the user’s actions on several platforms but also makes use of externally acquired data, such as data on credit card usage.

The core purpose of profiling is for the platforms to simply better understand their customers and develop their services. User profiles for individuals or groups allow targeting and personalisation of the offering based on user needs and preferences, and practical examples of making use of this knowledge include tailoring of services, price-discrimination, fraud detection and filtering of either services or users.

Pros and cons of profiling

From the user perspective, profiling is often discussed regarding problems that arise. Firstly, the data collected is largely from sources other than the individuals themselves, and the whole process of information gathering and processing is often a non-transparent activity. The user may thus have little or no knowledge of what is being known and recorded of them or how their user profile data is analysed. Secondly, how profiles are made use of in platforms, as well as how this data may be redistributed and sold onwards, is a concern. Discrimination may apply not only to needs-based tailoring of service offering and pricing but extend into ethically questionable decisions based on income, ethnicity, religion, age, location, the circle of friends, gender, etc. It should also be acknowledged that profiling may lead to misjudgement and faulty conclusions, and it may be impossible for the user to correct and escape such situations. The third and most serious problem area with profiling is when data and information on users is applied in harmful and malicious ways. This involves, for example, intentionally narrowing down options and exposure of the user to information or services or aggressively manipulating, shaping and influencing user behaviour. In practice this can mean filter bubbles, fake news, exclusion, political propaganda, etc. And in fact, the very idea of everything we do online or offline being recorded and corporations and governments being able to access this information can be pretty intimidating. Let alone the risk of this information being hacked and used for criminal purposes. Add advanced data analytics and artificial intelligence to the equation and the threats seem even less manageable.

However, profiling can and should rather be a virtuous cycle that allows platforms to create more relevant services and tailor personalised, or even hyper-personalised, content. This means a smooth “customer journey” with easy and timely access to whatever it is that an individual finds interesting or is in need of. Profiling may help you find compatible or interlinked products, reward you with personally tailored offers and for example allow services and pricing to be adjusted fairly to your lifestyle. In the future we’re also expecting behavioural analytics and psychological profiling to be used increasingly in anticipatory functions, for example to detect security, health or wellbeing deviations. These new application areas can be important not only for the individual but the society as a whole. Imagine fraud, terrorism or suicidal behaviours being tactfully addressed at early stages of emergence.

Where do we go from here?

Concerns raised over profiling are inducing actions in the public and private sectors respectively and in collaboration. A focal example in the topic of data management in Europe is the General Data Protection Regulation (GDPR) (EU) 2016/679 that will be applied in all European Union Member States from 25 May 2018 onwards. This regulation addresses the protection of natural persons with regard to the processing of personal data and on the free movement of such data. And even if launched as a European rather than a global initiative, the GDPR applies to all entities processing personal data of EU citizens, and many global players have in fact already claimed compliance in all their practices. Issues covered by the regulation include limiting the scope of personal data to be collected, the individual’s right to access data on them and detailed responsibilities for those processing personal data.

While the EU tries to manage the protection of personal data and thus bring transparency and fairness to profiling, the Chinese government is exploring a very different direction by being taking the lead in gradually introducing a Social Credit System. This model is at the moment being piloted, with the aim to establish the ultimate profiling effort of citizens regarding their economic and social status. Examples of the functioning of the credit system include using the data sourced from a multitude of surveillance sources to control citizens’ access to transport, schools, jobs or even dating websites based on their score.

Another type of initiative is the Finnish undertaking to build an alternative system empowering individuals to have an active role in defining the services and conditions under which their personal information is used. The IHAN (International Human Account Network) account system for data exchange, as promoted by the Finnish Innovation Fund Sitra, is designed analogously to the IBAN (International Bank Account Number) system used in banking. The aim with IHAN is to establish an ecosystem for fair human-driven data economy, at first starting in Finland and then extending to Europe and onwards. The plan entails creating common rules and concept for information exchange, and testing of the technical platform will be done together with pilots from areas of health, transport, agriculture, etc.

Selected articles and websites

Business Insider Nordic: China has started ranking citizens with a creepy ‘social credit’ system — here’s what you can do wrong, and the embarrassing, demeaning ways they can punish you
François Chollet: What worries me about AI
General Data Protection Regulation (EU) 2016/679 – EUR-Lex
Kirk Borne: Top Data Trends for 2018 – Part 1
Platform Value Now: Tackling fake news and misinformation in platforms
Sitra: Human-driven data economy
Wikipedia: Profiling (information science)
Wikipedia: Social Credit System
Wolfie Christl, Cracked Labs:‏ Corporate Surveillance in Everyday Life

Heidi Auvinen

Research Scientist VTT Technical Research Centre of Finland Ltd
Share this
TwitterFacebookEmailLinkedIn