January 8, 2020
Chief Information Officer, John Davison, talks to Information Age about how having a responsible attitude towards data and technology is key to maintaining customer trust.
Consumer trust can make or break a business, and the evolving technological landscape is raising new questions about how to demonstrate fairness. As companies turn to larger and new data sets to make pricing calculations and business decisions that directly affect consumers, concerns around inherent biases are escalating. It’s something I’ve seen in my own industry, insurance. With data science and new modelling techniques becoming integral to calculating premiums, there is a growing perception that customers could be treated unfairly and are therefore losing out financially.
This isn’t something we can afford to brush under the carpet. Across industries, we need to be asking ourselves how we can place fairness at the heart of this changing technological landscape. The answer lies in fostering responsible attitudes towards our data and our technology - achieved by using more sophisticated modelling techniques to help identify bias and assumptions, being transparent about how data is being used and educating customers about the benefits of these innovations.
Don’t assume
Across all sectors, from retail to healthcare, any company relying on data to calculate a price or influence the way customers are treated must make efforts to remove assumptions and biases from statistical modelling. It’s a trap that a business can easily fall into.
With every organisation possessing different historical data sets based on its unique customer history, a business might be tempted to account for a gap in information, perhaps a segment of the market not yet understood for example, with an assumption that isn’t rooted in hard statistics, but human bias. This is one example where we risk consumers being treated unfairly – for example, in the case of motor insurance, a demographic group paying higher premiums because of a lack of insurer history in a given segment of the market.
Of course, no business is going to create a data set that is entirely complete, but there are steps they can take to reduce the risk of bias and assumptions. First, they need to identify the limits of their current data sets, working to expose vulnerable areas where bias exists. They also need to make a difficult but necessary unbiased judgement on whether the statistical model is fair and accurate before using it.
They also need to invest in continuously expanding and evolving models. Given the rate of change in society more generally, data sets that are static and don’t move with changing customer bases are likely to become out dated and redundant and allow unfair assumptions to influence decisions.
As businesses become more data-driven organisations, relying on information to guide customer service innovations such as chatbots and calculate pricing, it’s vital they make every effort to identify and combat assumptions from statistical modelling. Otherwise, it will be difficult to ensure fair treatment, and feelings of customer trust will only decrease.
Honesty is the best policy
The right attitudes towards technology also involve being upfront with consumers about how their data is being used and, if applicable, how it feeds into pricing calculations. This doesn’t just benefit the public, but industry too, because a lot of ill-feeling is borne from a misunderstanding about how data is being used in different sectors.
In an insurance context, for example, young motorists often complain that they are unfairly treated, when they might actually be paying the same premium as a new 35-year-old driver who has also just passed their test, as the insurer’s data set shows that they present the same level of risk due to their lack of experience on the road.
In other words, in this example, it’s the level of experience and not discriminatory assumptions about age groups which may be motivating the pricing, contrary to perceptions. These misunderstandings arise from a failure to be fully open with customers, an area where the insurance industry – and others – are lagging. There is a need to highlight how these decisions are down to unbiased statistical modelling.
Arguably leading the way in transparency is the credit reference industry. By facilitating access to a customer’s credit score, it is helping people understand their individual financial position, enabling them to be better informed when planning their financial future and understand the various factors influencing their rating. Other sectors, insurance included, should follow their lead, empowering customers with information that matters to them, showing how different factors influence the prices they pay, and the difference certain behaviours would make – for example, demonstrating how cutting out bad driving habits could have a positive effect on premiums.
A transparent approach doesn’t just strengthen customer trust, but can lead to more meaningful relationships, immersing customers in the brand and showing how business objectives can align with them saving money.
Becoming an educator
Improving customer fairness doesn’t stop at being transparent when a customer makes a query or complains. It’s also about proactively educating customers about how technology can save them money. In energy, organisations such as Smart Energy GB are raising awareness of the potential of smart meters to make consumers more aware of their energy use and save on their bills, actively challenging misperceptions about why the technology is being deployed.
Similarly, in insurance, businesses can play an educational role by informing customers of how advances in counter-fraud technologies are ensuring that fewer fraudulent claims are being paid out. This not only reassures customers that insurers are doing everything they can to stop criminals benefitting at their expense but ensures that insurers’ resources are channelled towards genuine claims and ultimately driving the cost of premiums down.
The good news is that there will always be further technological advancement, making data easier to process and allowing greater accuracy. However, concerns around consumer trust are likely to remain. Ensuring fairness underpins our industries is a vital mitigation to these concerns and ultimately strengthens customer relationships. As we have seen, having the right culture, ethics and attitudes towards technology and data will be pivotal – not just in terms of boosting an organisation’s reputation, but ultimately impacting the bottom line in the form of customer retention and acquisition.
Consumer trust can make or break a business, and the evolving technological landscape is raising new questions about how to demonstrate fairness. As companies turn to larger and new data sets to make pricing calculations and business decisions that directly affect consumers, concerns around inherent biases are escalating. It’s something I’ve seen in my own industry, insurance. With data science and new modelling techniques becoming integral to calculating premiums, there is a growing perception that customers could be treated unfairly and are therefore losing out financially.
This isn’t something we can afford to brush under the carpet. Across industries, we need to be asking ourselves how we can place fairness at the heart of this changing technological landscape. The answer lies in fostering responsible attitudes towards our data and our technology - achieved by using more sophisticated modelling techniques to help identify bias and assumptions, being transparent about how data is being used and educating customers about the benefits of these innovations.
Don’t assume
Across all sectors, from retail to healthcare, any company relying on data to calculate a price or influence the way customers are treated must make efforts to remove assumptions and biases from statistical modelling. It’s a trap that a business can easily fall into.
With every organisation possessing different historical data sets based on its unique customer history, a business might be tempted to account for a gap in information, perhaps a segment of the market not yet understood for example, with an assumption that isn’t rooted in hard statistics, but human bias. This is one example where we risk consumers being treated unfairly – for example, in the case of motor insurance, a demographic group paying higher premiums because of a lack of insurer history in a given segment of the market.
Of course, no business is going to create a data set that is entirely complete, but there are steps they can take to reduce the risk of bias and assumptions. First, they need to identify the limits of their current data sets, working to expose vulnerable areas where bias exists. They also need to make a difficult but necessary unbiased judgement on whether the statistical model is fair and accurate before using it.
They also need to invest in continuously expanding and evolving models. Given the rate of change in society more generally, data sets that are static and don’t move with changing customer bases are likely to become out dated and redundant and allow unfair assumptions to influence decisions.
As businesses become more data-driven organisations, relying on information to guide customer service innovations such as chatbots and calculate pricing, it’s vital they make every effort to identify and combat assumptions from statistical modelling. Otherwise, it will be difficult to ensure fair treatment, and feelings of customer trust will only decrease.
Honesty is the best policy
The right attitudes towards technology also involve being upfront with consumers about how their data is being used and, if applicable, how it feeds into pricing calculations. This doesn’t just benefit the public, but industry too, because a lot of ill-feeling is borne from a misunderstanding about how data is being used in different sectors.
In an insurance context, for example, young motorists often complain that they are unfairly treated, when they might actually be paying the same premium as a new 35-year-old driver who has also just passed their test, as the insurer’s data set shows that they present the same level of risk due to their lack of experience on the road.
In other words, in this example, it’s the level of experience and not discriminatory assumptions about age groups which may be motivating the pricing, contrary to perceptions. These misunderstandings arise from a failure to be fully open with customers, an area where the insurance industry – and others – are lagging. There is a need to highlight how these decisions are down to unbiased statistical modelling.
Arguably leading the way in transparency is the credit reference industry. By facilitating access to a customer’s credit score, it is helping people understand their individual financial position, enabling them to be better informed when planning their financial future and understand the various factors influencing their rating. Other sectors, insurance included, should follow their lead, empowering customers with information that matters to them, showing how different factors influence the prices they pay, and the difference certain behaviours would make – for example, demonstrating how cutting out bad driving habits could have a positive effect on premiums.
A transparent approach doesn’t just strengthen customer trust, but can lead to more meaningful relationships, immersing customers in the brand and showing how business objectives can align with them saving money.
Becoming an educator
Improving customer fairness doesn’t stop at being transparent when a customer makes a query or complains. It’s also about proactively educating customers about how technology can save them money. In energy, organisations such as Smart Energy GB are raising awareness of the potential of smart meters to make consumers more aware of their energy use and save on their bills, actively challenging misperceptions about why the technology is being deployed.
Similarly, in insurance, businesses can play an educational role by informing customers of how advances in counter-fraud technologies are ensuring that fewer fraudulent claims are being paid out. This not only reassures customers that insurers are doing everything they can to stop criminals benefitting at their expense but ensures that insurers’ resources are channelled towards genuine claims and ultimately driving the cost of premiums down.
The good news is that there will always be further technological advancement, making data easier to process and allowing greater accuracy. However, concerns around consumer trust are likely to remain. Ensuring fairness underpins our industries is a vital mitigation to these concerns and ultimately strengthens customer relationships. As we have seen, having the right culture, ethics and attitudes towards technology and data will be pivotal – not just in terms of boosting an organisation’s reputation, but ultimately impacting the bottom line in the form of customer retention and acquisition.