Digital transparency on the job: playing with data

Prof. Alexander Pretschner, Chairman of the bidt Board of Directors and a computer scientist, heads the interdisciplinary project “Inverse Transparency” together with the social scientist Prof. Andreas Boes and the economist Prof. Thomas Hess, both members of the bidt Board of Directors. In the interview, they explain how cultures of innovation can be thought together with data protection and how ever more extensive digital control of employees can be prevented.

In your project, you deal with data that accumulates during work – what is that, for example?

Thomas Hess: Today, as a result of digitalisation, companies automatically collect a lot of data about what employees do. This has been known for some time, for example in the insurance industry, where a lot of data is generated when processing claims. Such automatic recording, which was previously more technically driven in order to optimise certain systems, now exists in many areas.

We examine this using the software industry as an example. The tools that software developers use contain a lot of records about who did what and when. So it’s all about data about what employees do.

Andreas Boes: Data is generated automatically today as a by-product of working. There is no camera anywhere that records what someone does, how often and how long they are there. But anyone who works with digital systems – and most people do by now – is constantly generating data.

What can this data be used for?

Thomas Hess: Intuitively, one would think that it is clear what is happening: Supervisors who have so much data about their employees will switch to micro-control, that is, they will control every little activity of all the employees for whom they are responsible. But that would be exactly the wrong thing to do for two reasons. First, from a management perspective, it is not said that this is the better kind of leadership. For example, it would not motivate employees. And the second is that there is no acceptance for it. Employees would certainly find a way around such a system or, if labour markets were employee-friendly, they would simply leave and change employers.

Recently, the company Zalando landed in the headlines for allegedly monitoring its employees with software. Does that show how things should not be done?

Andreas Boes: That is a very interesting negative example. Basically, a sensible tool, namely 360-degree feedback, was misused to entrench a snitch culture in which everyone snitches on everyone else – at least if the study that was published on this is to be believed.

Thomas Hess: Such a system is not sustainable, it is exactly the opposite of a culture of trust. Leadership will never work like that, especially in the future.

Andreas Boes: Our opinion is that a sensible work process cannot be achieved today without a culture of trust.

Technology makes it possible to control people down to the deepest pores of their performance. One answer to this could be to radically protect data that accumulates in the work process. However, this means losing the associated opportunities to use it sensibly, for example for innovations.

Conversely, if all data were released, there would be a danger that people would not get involved in working with it because they would be permanently afraid that the information could be used against them. In our project, we are trying to find a solution to this apparent design dilemma in interdisciplinary cooperation.

How do you want to ensure that in your project?

Thomas Hess: The basic idea is to create a win-win situation. That’s why we are looking for a solution for the use of this data from which both sides have something. And we want to achieve this through inverse transparency.

What do you mean by “inverse transparency”?

Andreas Boes: An employment relationship is always asymmetrical. The right of direction determines that two equals do not meet, but one side has more power. The danger now is that the data that is created today ends up with the management and nothing happens except that it gets even more power.

In contrast, we put forward the idea of inverse transparency: employees see what is seen. They have insight into what data is being used by them and for what purpose. This way, they can be sure that nothing is happening behind their backs. The principle behind this is “Watch the Watcher”.

How can this be achieved technically?

Alexander Pretschner: We are developing a tool that can log the use of data in a tamper-proof way. In the project, we are also building a tool with which the collected data can be aggregated and displayed, so we have control over the logging.

From a technical point of view, there are a number of ways to access employee data: One can directly access the database in which the data on work processes are stored. You can access the data via aggregation tools, which are designed to add value to the data collected. You can export Excel tables that can then be used somewhere, and you can photograph the screen.

We have been working on approaches of so-called distributed data usage control for the last ten years, which can be used to control the use of data after it has been released. In this project, we are now focusing on the aggregator and the database, as well as tamper-proof logs that record access.

There are still a number of exciting challenges: How do we deal with large amounts of data? And how do we make sure that the user understands what the limits of this approach are?

What does access to this data mean for managers?

Andreas Boes: The prerequisite for using the data in a meaningful way is a culture of trust. That is why it is so important that we approach the topic in an interdisciplinary way. Thomas Hess deals with questions of leadership in data-mediated structures. And I approach the topic from the staff side: how do we get staff to actively use this data? An important element is rules that define how the game with data works.

Thomas Hess: Rights and rules are needed for the technical side, but also for the management relationship.

Alexander Pretschner: From a technical perspective, we can define and enforce in defined contexts who is allowed to access the data, when, for how long and under what circumstances, also when the data is deleted – but of course these rules have to be defined by people.

Can work and performance always be represented in data at all? How does that work for so-called knowledge workers or in creative professions?

Andreas Boes: For simple work, for example, the strokes are measured.

Thomas Hess: … or in industrial production, how many parts are manufactured.

Andreas Boes: Software development, on the other hand, is basically a highly creative activity that produces very complex and also innovative products. We are working on the project with Software AG, one of the largest software manufacturers in Germany.

The company is very conscious of the fact that it does not measure everything it could measure. Instead, it tries to capture in data transparency what is really important. For example, how a problem was solved. These experiences are then shared with all employees worldwide.

Thomas Hess: At management level, it is interesting that the data is not used to monitor individual progress. You could look at the software and see how many lines someone has written and think that someone who has done a lot of programming is productive. Or it could be determined by the result: Maybe someone has only written three modules, but they are the better ones. But such a control would not make sense from the company’s perspective. It would not lead to the management receiving input for innovations.

The keyword innovations has already been mentioned several times: How can this data be used for this purpose?

Alexander Pretschner: There is a research trend called code intelligence. In the past, we computer scientists tried to draw conclusions from the code alone, for example, where there are particularly error-prone areas or which pieces of code are particularly difficult to maintain – using metrics with which we recorded structural features of the code. That didn’t work very well. But if you add contextual knowledge from the development process, it works much better. We can then sort test cases, for example, so that those tests that are most likely to find errors are executed first, so that developers get feedback quickly.

I believe that development-related data can have an extremely positive influence on software development – as long as it is used in a human-friendly way.

Andreas Boes: So it’s not just performance-related data that accumulates. A lot of data is increasingly important for people to be able to work together sensibly. For example, in some companies, all communication runs in systems similar to the one we know from Facebook. And modern companies in particular use data to create transparency.

There are programmes where colleagues can see how far along the work process they want to connect to is. For example, they can plan when it makes sense to start.

In the new data-based value creation systems, innovation and value creation go hand in hand. Companies like Facebook, Google or Spotify are constantly expanding and improving their product. They need data-mediated transparency so that employees can innovate. They can see in real time, for example, how a customer is using their product and then think about how it can be improved.

Thomas Hess: Our project starts with the changed management processes, but the basic idea goes beyond that: that ultimately a win-win situation is found when you manage based on data. This can also be transferred to other areas. Innovation is certainly the most important area from an entrepreneurial point of view, but there are other fields, for example occupational safety.

Andreas Boes: We will start in 2020 with the company practice laboratory at Software AG. In the third year, technology and the tools will be transferred to a larger group of companies. Our goal at the end is to have a case-tested solution for handling data in digital work processes.

Alexander Pretschner: In the long term, the approach should of course not only be used for software engineers, but also for managers.

Prof. Andreas Boes, Prof. Thomas Hess and Prof. Alexander Pretschner are jointly leading the project “Inverse Transparency: Designing Participatory Approaches to Data Sovereignty in the Digital World of Work”, funded by the Federal Ministry of Education and Research, which started in 2019. For this, they are working together with the company Software AG.

---