The Problem with Putting Your Trust in Data

As an avid football fan, I was keenly interested when earlier this year I heard about Manchester City’s latest signing. Someone who is considered essential to the future dominance of that club. Imagine my surprise to learn that it was a data scientist who will use data analytics to improve the fitness and skills of the team, and support recruitment by helping to identify promising players and perform analysis of their performances. It seems that across all organizations, even in “the beautiful game”, data-driven innovation is now essential.

So, it is worrying that when I am working with large established organizations on their digital transformation strategies, how quickly the excitement dissipates when the conversation turns to fundamental questions about data. As digitization of manual steps turns to transformation of processes and practices, the data-driven nature of many digital transformations requires questions to be asked about how to address the responsibilities that come with the ownership, stewardship, and maintenance efforts associated with that data.

Much of the discussion concerns how to balance the opportunities and risks of data-driven digital transformation. In this context, some people have described data as the new oil. While others see data as a new currency that has its own intrinsic value or can be traded and exchanged for goods and services. In both cases, while the metaphors may be over-played, they emphasize that data is a key asset that must be collected, curated, managed, and shared. They also highlight a broader set of obligations.

As with all assets, the ownership and control of data is a central issue. It brings many opportunities for creating value. But also carries important liabilities with respect to maintaining its accuracy, ensuring it is not misused, and protecting its providers’ privacy and confidentiality. This is a difficult issue in the case of data that is internal and proprietary to a single organization. It is yet more complex when multiple parties are involved. Particularly when gathered from multiple sources and shared by diverse stakeholders, how should data governance be handled?

For many people, the answer lies in data trusts. Mirroring the idea of trusts used in law as mechanisms to manage assets such as property and investments, they establish the role of a trustee to be responsible for decision making about data assets on behalf of the various trustors. By working on their behalf, they balance the potentially competing concerns to satisfy the needs of all stakeholders and may serve different purposes: they can be used across commercial consortia, to create common data sharing with appropriate protection, or to open up data for use in non-commercial activity.

With this range of applications, data trusts are gaining support. Much has been written about their rolepilots have been explored, and various models of data trusts proposed. And although many variants exist, at their core they can be defined very simply as legal structures that provide independent stewardship of data. In practice, however, there are several different elements that must be provided. In the work of the Open Data Institute (ODI), they highlight several characteristics essential for the success of data trusts:

  • A clearly defined purpose.
  • An effective legal structure (including trustors, trustees with fiduciary duties and beneficiaries).
  • A defined set of rights and duties over stewarded data.
  • A well-specified decision-making process.
  • A description of how benefits are shared by all stakeholders.
  • A sustainable funding model.

But data trusts are not the only game in town. This week, a report produced by the Ada Lovelace Institute with the AI Council has highlighted three different data stewarding mechanisms that support shared ownership of data for the public good. Alongside data trusts, they describe the features of data cooperatives and corporate and contractual models. While they share many of the same goals, the report distinguishes when each of these data stewardship approaches may be appropriate:

  • Data trusts emphasize a highly participatory approach when all the stakeholders work together to define the terms of the trust and support the trustees in their operational tasks.
  • Data cooperatives are useful when a collection of stakeholders pool data and repurpose that data for different uses supported by those involved.
  • Corporate and contractual models help to build trust between two or more parties when sharing data to support specific business or community objectives.

Each has their uses. However, it is important to emphasize that in all three cases the focus is placed on the trustworthy and responsible approach to the use and management of data. They create opportunities for sharing data within contexts that offer assurance to all stakeholders that governance is provided in their best interests.

And in the end, it is that alignment of incentives that is likely to make the difference in bringing data sharing to life. Maybe this is the way we might see Manchester City’s new chief data scientist sharing data on player performance with the likes of Liverpool and Arsenal as a way to understand more about player’s long term health issues, how to reduce stress injuries, and where they focus to maximize their performance. Or there again, maybe some rivalries simply run far too deep.

Source: AWB Digital Economy Dispatch #26

Alan Brown

Alan W. Brown is Professor in Digital Economy at the University of Exeter Business School where he co-leads the Initiative in Digital Economy at Exeter (INDEX). Alan’s research is focused on agile approaches to business transformation, and the relationship between technology innovation and business innovation in today’s rapidly-evolving digital economy. After receiving a PhD in Computational Science at the University of Newcastle-upon-Tyne, Alan spent almost 2 decades in the USA in commercial high-tech companies leading R&D teams, building leading-edge solutions, and driving innovation in software product delivery. He then spent 5 years in Madrid leading enterprise strategy as European CTO for IBM’s Software group. Most recently Alan co-founded the Surrey Centre for the Digital Economy (CoDE) at the University of Surrey where he led research initiatives in 4 EPSRC-funded research projects.

Explore more content

Improving in the moment

Seeing the opportunity
Read more

Being an Agile developer

Knowledge workers are key to Agile <
Read more

Unit and Integration Testing Overview

The two important types of testing <
Read more
Contact Us