Researchers encourage an interpretation-based model in information management

For the past four years, two research teams, Professor Jukka Heikkilä and PhD Marikka Heikkilä and Professor Tomi Dahlberg and a PhD student Tiina Nikkala from Turku School of Economics, have made theoretical studies of how information should be managed in modern organisations, specifically where information is diverse and scattered across different data networks and information assets. The study is part of the Digile D2I (Data to Intelligence) research programme (http://www.datatointelligence.fi ). Recent results show that data models should be separately considered otherwise organisations could face large problems with their information management.

The amount of information is increasing exponentially. Public organisations and companies need to consider, what is data; how can it be managed; who is allowed to use it; and how can data be combined to make it beneficial for both current and future operations. Incorrect information can lead to false conclusions that result not only in lost business, but in extreme circumstances can lead to severe social ramifications, such as with data pertaining to cancer treatments or terrorism risk assessments.

Information in three dimensions
The researchers have defined three dimensions for information: data, meta data and meta meta data. The first dimension (data) provides the core information, with the second dimension then providing its descriptive meaning (meta data). The third dimension (meta meta data) forms the interpretative layer on top of meta data, explaining under which circumstances information can be used and on what terms. On a practical level, metadata is categorised as technical (meta), informative and interpretative (meta-meta).

At the beginning of the D2I research a survey was conducted to find out how participating organisations distribute information throughout their operations, who owns the information, how it is used and how it is transmitted. In particular, the objective was to identify the current state of interpretative meta meta data in organisations.

“There are requirements for information, such as, whether it can be used; is it protected; what should be done with the information to make it usable; and under what conditions can it be used. Meta meta data of this nature was not found among the organisations surveyed. It turns out that we have more work ahead of us than we had expected,” explains Professor Jukka Heikkilä.

Information has different meanings in different situations
Master data management, one of the main tasks in information management, was also covered in the study. A majority of organisations manage master data according to a golden record approach. The idea behind this approach, originating from enterprise resource planning systems (e.g. SAP), is to have only one entity of master data, which is forced into a specific format. For example, a customer is always a customer, even though it could have different meanings in different contexts within the same organisation. Researchers consider this an error in thinking:

“It just doesn’t work. Not just because entities have different meanings in different situations, but also because over time what we want from information changes. Also, information classifications can evolve,” states Professor Tomi Dahlberg. Tomi Dahlberg has followed and studied this phenomenon for many years.

To cite an example, Tomi Dahlberg refers to financial management operating in a restricted and regulated environment. Analyses are often reliable until a new company is acquired. The two companies have different ways of recording things such as depreciation, recording methods can change or the financial system could be replaced. Similarly, stocks valued in the balance sheet can have many prices for the same product (e.g. a component or spare part). Reporting no longer provides reliable information for the management about the overall operational status, not to mention future predictions. Uncertainty increases when the time comes to consider the integration of different product and customer information.

 Interpretation maximises benefits from existing systems

During D2I it became clear that organisations, without exception, have a need to complement and enrich existing information with a local perspective. This is the opposite of the golden record thinking, where it is agreed that a centralised box includes one and the same data that everybody must use. 

 In the real world large organisations have hundreds of information systems that enable operations. Millions, if not billions have been invested in different systems and renewing them at once is not a realistic option. The smartest way according to Tomi Dahlberg is to leave existing systems as such and, with a step by step approach, combine information assets by using meta meta data. The structure of information is generated by its true use and utility.

There is a small group of highly advanced and skilled companies in Finland who can exploit information in their business. These companies can be found in the gaming, machine shop and telecommunication industries.  Most Finnish companies are lagging in their way of thinking. Professor Dahlberg is concerned by the situation:

“Most companies just carry on like before. This is a concern. People talk about digitalisation, but don’t really understand how to utilise information for business and use it to generate new business.”

“Within public administration the problem is twice as bad,” continues Jukka Heikkilä. “On the one hand Finland is advanced in utilising areas such as electronic medicine prescriptions, despite it taking 17 years to build and deploy the system. However, on the other hand, the IT systems in healthcare, delivered by a few large IT system providers, are still not able to interact with each other.”

The need to combine data from different sources increases
The future will present an even wider range of data sources, like big data and the Internet of Things. At the same time social media produces wild and free data, which, when analysed properly, can provide answers to different phenomena, like a sudden drop in sales.

The fast transition to cloud services among organisations means less investment in their own data centres and equipment. The challenge of information management in organisations is to understand, both technically and meaningfully, what information is where and how to wisely combine all data sources.

“Information management needs new tools to control scattered information assets, and above all how to interpret and report on these dispersed pools of information,” Jukka Heikkilä says.

Surprising and important findings

The most relevant finding of the study, according to Tomi Dahlberg, is the fact that information must be managed from a perspective that understands its meaning in different situations where data is generated, used and reported. In addition, it is important to be able to monitor changes in the status and describe information in such a manner that things become clear.

“During the study, the lack of meta meta data has been most surprising.”

Jukka Heikkilä points out that too often the value that one attributes to information leads to the prohibition of its use. There is a lack of appreciation and understanding about the value of sharing information – sharing information is not the relinquishing of an asset but instead the opportunity to discover something new. This happens when privacy is overprotected. According to researcher Marikka Heikkilä, there are lots of rules and legislation related to public administration information in particular regarding what can and cannot be transmitted. Information systems should describe reality, but in some cases this is not allowed. For example, in heath care there are situations when sharing data is not allowed even when treatment would require this.

Practise supports theory
According to Tomi Dahlberg, the core of the issue is relatively simple: should it be assumed that the meaning of data is always the same or should there be provision for conceptual interpretation. The idea that claims data everywhere is one and same, is bewildering! Traditions are very strong, but smart people don’t consider this correct when trying to pursue a cohesive and uniform model for data.

“With information increasing exponentially we must correct our thinking now, not in 10 years,” says Dahlberg.

This is a common problem for everyone. The new way of thinking has been applied in practice among a few companies in Finland. Apart from these few companies researchers are not aware if challenges have been resolved anywhere else in the world.

“Ineo is onto a solution for how data models and systems can be combined. Now we know that our theoretical studies can be applied in practice,” tells Jukka Heikkilä.

Ineo has participated in the D2I research and empirically implemented conceptual interpretation of data as a part of the research and among its customer projects.

In the future the team plans to engage with the following: what is the service or organisation in healthcare that can guarantee that the meaning of information is understood in its various applied situations, so that information will describe timely changes in status and that reporting based on information is understandable. These are some of the questions PhD student Tiina Nokkala examines in her doctoral thesis.