What is data discrimination?
The idea of Data Discrimination – alternatively called ‘discrimination by algorithm’ – is proven to exist, defined as “bias that occurs when predefined data types or data sources are intentionally or unintentionally treated differently than others.”
For commercial and marketing departments who rely on segmentation on their websites and in their databases, these rules are centred on the very idea of inclusion and exclusion. Whether conscious or not, manual interventions in using data belie institutional discrimination – from companies that send Mother’s Day email reminders just to men on their database to digital advertising campaigns for luxury products that only target homeowners of properties worth at least £1 million.
Data practices in marketing
Technology – and how it uses data – is often skewed towards the ‘average’ or a dominant group. The whole concept of A/B testing reinforces this, with average results (open rate; click through rate to name just two) being applied across the remaining audience. Even if you’re running A/B testing on carefully segmented mailing lists, the ultimate definition of ‘success’ is made by data decisions based on averages.
Risks of data-led ‘averages’
By working with averages, you are prioritising a generalised ‘ideal’ customer who, at best, might reflect just some of a user’s priorities. Making the wrong assumption about a user on your website can result in them turning to a competitor. If they feel you are not speaking for them, they will look elsewhere for their ‘tribe’. Algorithms for data-decisions should be consistently self-learning: by utilising multi-faceted AI, your website can react – and interact – in real-time with each individual visitor. The ability to scale conversations to each individual in real-time is only possible with AI; manual interventions simply cannot serve thousands of users at the same time, each in a different way that speaks directly to the individual. By no longer relying on averages, you remove the risk that algorithms are trained on one group and inappropriately used to measure another.
How data can be used to value diversity
Advanced technologies present opportunities for marketers to value diversity and, importantly for businesses, can help you improve your business results. By using tools such as ours – which work on multi-variant, advanced algorithms to shape your digital presentation in real-time to each individual user – you are able to move away from the dominant approach of valuing averages, towards a tailored, individual-centric approach. This can apply to everything from what product choices are shown to the images a user sees. To speak to the diverse range of individuals your website may have, you need real-time data processing to facilitate immediacy.
Opportunities for effective data use
It’s by acknowledging the differences your customers may have that you can prioritise freshness in search over best sellers. But how do you know which user values what experience? Will User A react positively to a picture of a family on holiday when researching summer breaks, or do they want to see a group of friends lounging by a pool?
Reaching a new level of personal relevance and service quality in the online travel shopping experience requires input from your user, which may come from direct user input or be deduced from their online behaviour. According to a 2018 survey, 51% of global consumers will provide their data if there is a clear benefit to them.
As well as meeting GDPR and local regulations, ensuring you have a transparent data policy for users to understand how you store and use their information is key to encouraging them to give their consent. The same survey also noted “consumers want more transparency (86%) and control (83%) when it comes to their data in order to build these levels of trust.” How you gather and use data is key to building trust. Place the user at the forefront of everything you do. Make it easy for them to view their profiles and update information: a loyal user may have once wanted a couple’s break to Iceland, but now wants a family holiday somewhere sunny with a kids’ club.
Put the user in control
Let users easily provide information they are comfortable sharing. By putting the user in control, with the ability to switch on or off their information, they can gauge the impact of personalisation on their own experience with your site. In the same way a user may try a browser’s incognito mode to see if the pricing changes, give them the chance to see how providing their data vs not providing their data enhances their search experience.
Think outside the box with how you measure success. Our CEO Andy recently spoke about the importance of measuring success, and it’s only by putting the user at the heart of everything you do – including relinquishing control over what they do and don’t tell you, whilst giving them reasons to provide their personal information – that you can reach a new level of personal relevance and service quality in your online travel shopping experience.