Data is no longer simply part of a business’ arsenal of tools: it is the very source of its power and innovation. But with the old adage of power and responsibility, executive teams are still working to understand the complexities and challenges that data still presents. Namely, bias. Not only does this have grave implications for end users, seen bias in data also affects the trust we place in those brands and in their solutions or products. And as customers become ever more aware of their own data and its use, companies are beginning to realize just how fragile that implicit trust is.
In Data Visionaries: Challenges and Technologies, moderator Emma Sinden put it to data leaders Doug Laney, Data Strategy lead of West Monroe, Daniel Jeavons, Data Science lead of Shell and Maurice Coyle, Chief Data Scientist at Truata, asking how those trends are interplaying, and what businesses and their Chief Data Officers should be considering to maintain that trust alongside innovation.
The value of trust
“Trust is so important,” said Maurice Coyle, “in so many industries it can be considered an existential requirement. More people are becoming aware of data collection and use, and being proactive in demanding evidence of how businesses are using their data.
“Companies ought to be transparent with their practices and give control back to the customer, in order to maintain or generate trust.”
Trusta recently conducted a study, asking individuals whether respected use of data was a factor in their purchasing decision. Over 70 percent of respondents said they would be willing to leave a brand if their data wasn’t respected. That showed, concluded Coyle, customers would vote with their feet.
Giving control to customers, however, might mean different things for different companies. Some data leaders believe customers won’t want to navigate comlex pages or documents in order to satisfy their demands for data transparency and so want to do it themselves by designing simpler data architecture. Others, however, consider this a contradiction in terms. They believe customers want to have more and more direct control and any methods to reduce that may be misinterpreted.
Education, it seems, is the way to reduce misinterpretations.
“We’ve talked for years about the importance of a good data governance function to ensure data is correct, complete and managed properly,” said Doug Laney, “but hardly anything is heard on analytics governance. Someone needs to vet your data models across the organization, peer-review style, as we’re on the cusp of self-service analytics”
This practice will also incorporate data literacy, he added, that should revolve around ethics, bias and logical fallacies, which nearly always creep into models. This education and awareness would help the organization design better data strategies with the customer in mind.
Clearing the bias fog
Shell’s global data science lead Daniel Jeavons agreed and detailed the changes he’s put in place to offset those fallacies and human errors.
“We created a global discipline head, a scientific lead for the data practice who’s responsible for setting science standards, curriculums, training and programs of assurance so teams are held to account,” said Jeavons.
He also reported a new tool that is helping his team focus harder on how they code. The “inner source program” enables everyone to see everyone else’s codes at all times.
“People are far more likely to write better codes now knowing their boss can access and read them at any time,” he noted, wryly.
It’s another form of transparency that the roundtable believed was necessary to protect businesses and end users alike. Many businesses have multi-layered health and safety procedures for physical work, so why not replicate that for the digital, data landscape? After all, the central challenge for the ‘Data Visionary’ will not be how to collect more data or how to interpret its information, it will be how and when to use that data using a refined methodology of peer review, ethics, business strategy and trust-based decision making.
This roundtable is brought to you in partnership with Truata, a leading provider of privacy-enhanced data solutions. Specializing in privacy risk assessment, de-identification and true anonymization of data, Truata’s suite of proprietary software solutions and services enables businesses to unlock powerful insights while ensuring that they comply with the highest level of global data protection standards.