Big Data is a necessary condition for automation and robotization in business. This usually relates to processes and/or fields that require intensive human effort. The manufacturing sector can provide countless examples of this, but in the context of financial services, robotization focuses mainly on the work of intermediaries. This includes the automation of mortgage advice, the mediation involved in taking out insurance or performing property valuations.
The obligation of data minimization
The collective drive to automate these processes has led to the accumulation of vast amounts of sensitive consumer data. Much of this data was collected when the General Data Protection Regulation (AVG) and the Facebook-Cambridge Analytica scandal were still far off in the future. Privacy, in short, was still an under-researched topic when the principles surrounding Big Data collection were taking shape. In an era of legally enshrined consumer consent and the obligation of data minimization, Big Data has suddenly become a risk for many organizations. As evidenced by the high-profile GDPR-related fines for Booking.com, the Dutch Employee Insurance Agency (UWV) and the Bureau Krediet Registratie (BKR) Foundation among others, sometimes storing less data is better. After all, data no longer present cannot be leaked, misused or stolen.
Limited return on investment
In addition to the risks posed by the abundance of data, many organizations are particularly concerned about the cost of efficiently managing data. In order to use Big Data for automation, the quality, consistency and completeness of this data is of great importance. The latter is demonstrated, for example, in research by the American Stanford professor Andrew Ng, who specializes in Artificial Intelligence and Machine Learning. According to the professor, the costs of identifying, cleaning and structuring so-called ‘bad data’ are particularly high. Of all the data science work required to automate business processes using Big Data, he estimates that 80% of the work is spent on cleaning and preparing the necessary data. In short, to get a higher return of investment on data collection, the quality at intake must be improved. This is where Good Data will play an important role, according to NG.
Good Data is information that harmoniously brings together the needs of the triangle of consumers, businesses and regulators. Where consumers seek control over their own data, the business community seeks as much insight as possible into consumer purchasing patterns, and the regulator monitors compliance with laws and regulations such as the GDPR or Wwft. This harmony is currently lacking and this is largely because the collection frenzy surrounding consumer data is ahead of the legislation that seeks to regulate it. In the financial sector in particular, this is leading to problems. For example, the research report ‘Regulatory Technology on the Rise’ by Dutch fintech Hyarchis shows that only 15% of financial institutions believe they have a grip on their historical customer data.
Asked about the research, Adriaan Hoogduijn, COO of Hyarchis, states: “The development of Big Data can be compared to the historical development of the agricultural sector. As science and the resulting technology have developed, agriculture has become increasingly efficient. From meeting the immediate needs of the local community, agriculture has rapidly scaled up to become a massive bio-industry. While this industry maximizes the amount of food available, it does not contribute positively to global health. In a similar way that agriculture is becoming more ecological, data collection will also move toward a healthy middle ground between all stakeholders. Big Data has reached the limits of healthy scalability and the move towards Good Data has become unavoidable as a result.”