Lead User Researcher, Copywriter, Visual design and infographics
The General Data Protection Regulation took effect on the 25th May 2018 in the EU. On that day, every organisation who held personal data on individuals had to ask for permission again to keep their data.
They say that data is the new oil. Artificial intelligence (AI), in particular machine learning (ML), has been adopted by many companies as well as governments across the world. The cornerstone of ML however, is based on the availability, quality and diversity of data used to train the models. Without a diverse and vast set of data, these AIs may become biased, or may even harm those groups who are not properly represented.
On the other hand, we see an increased awareness by the everyday person and well as governments about the user of personal data. With new European legislative changes such as GDPR, as well as the public's reaction to the Facebook/Cambridge Analytica scandal, how does the general public perceive the value of their data?
We conducted research using an online survey and other qualitative methods to investigate from the user’s perspective, how they decide on what data to give to companies. Here are our top level findings:
We found that trust and user value are essential to whether or not a user hands over data. However, trust is more important than the customer value (perk), therefore gaining trust from users is vital.
Eiser and White (2008) theorised that trust can be categorised into baseline trust (built over years of advertising, word of mouth and high street presence) or marginal (the trust gained and loss after an event, such as a hack, a rebrand etc.). We introduce the two types of trust, how it is acquired and lost, and ways that newer challenger companies can growth-hack baseline trust.
Traditional baseline trust requires a historical presence. We found a few ways that a company can hack trust growth: Leveraging the right people, being honest and transparent and partnering with the right experts.
We found differences in mental models of trust among different industries. First, people trust the government most to keep their data safe, followed by regulated sectors (i.e. banks, telcos, airlines) then non-regulated sectors. We also found that the criteria people use to determine whether a company is trusted varies across verticals.
Aims:
01. Secondary Research
To gain a solid understand of the trust/data economic landscape, secondary research was used to gain a general understanding of the landscape and identify and evaluate ideas put forward by other researchers. Most of the research was industry specific, so we had to pull together general findings across industries to come to our conclusions.
Key finding one:
Key finding two:
Baseline Trust is the intrinsic trust in a brand or a company based on its past public relations and brand identity. New start ups have less baseline trust and traditional brands and institutions have higher levels of baseline trust
Marginal Trust is event related. It’s the trust built or lost from specific events such as good customer service experiences or a data breach Over time, positive marginal trust turns into baseline trust. Similarly the impact of events that damage marginal trust can start to erode established baseline trust.
02. Online Surveys
Using Survey Monkey, an online survey was completed by 130+ participants in the UK to understand their attitudes towards data sharing and how trust is perceived. The survey was also used to validate the desk research.
Using Survey Monkey, an online survey was completed by 130+ participants in the UK to understand their attitudes towards data sharing and how trust is perceived. The survey was also used to validate the desk research.
03. Qualitative Study
Based on the findings from they survey, 20 of the survey participants were selected to complete a face to face study. Participants were asked to complete three separate tasks under observation.
To understand the relative importance of 50+ different data types such as location tracking when app is opened, dietary requirements, and health records, participants were asked to sort different data types into high, mid and low sensitivity.
By asking them to vocalise their rationale, more in depth insight was gained on the mental models in categorisation.
Half of our qualitative research participants were presented with 3 banks (regulated industry) to choose from; a high street bank that they have had all their life, a challenger app type bank (similar to Monzo and Startling) and a “piggy back trust” bank account (i.e. Bank of Ireland piggybacking on the Post Office’s heritage to rebrand their banking products in England).
The other half were presented with 3 food/grocery companies: A high street supermarket Woods (similar to Tesco and Sainsbury’s in the UK),a meal kit subscription service (startup) and an online only grocery delivery aggregator “piggy backing” on a tech giant (similar to Amazon Prime Fresh and mysupermarket).
Both groups were asked the same question: which company would you choose to use as your primary banking/grocery provider? It’s important to keep in mind that the choice was made as the primary service provider. Many of our participants voiced that they would have multiple bank accounts and shop at multiple outlets, but if they had to pick one, these would be their choices.
After the initial choice, participants were presented with a variety of perks (i.e. 5% cashback) and asked to personalise the product offering of the company they did not choose. The question “what would it take for you to switch to this provider?” was asked and a custom product was designed by the participant.
Each of the product perks did not cost money, but a data price shown (i.e 20% off purchases over £100 would cost someone their email address and browsing history). Participants paid for the perks using data from the data wallet they created in Task One for perks. The prices were also up for negotiation, and researchers were allowed to change the price and or the perk, so as to gain a deeper understanding of the rationale behind the choices.