A digital extended family instead of a digital twin

A window is a window. Digitally, however, the same window can exist in several places at once: as a component in a BIM model, an element requiring maintenance in a CAFM system, a commercially relevant asset in an ERP system or an entry in other databases.

Lars Oberwinter encounters these system landscapes in his daily work. He is Managing Director of Plandata GmbH, which advises companies on BIM strategies, digital standards, data structures and connecting digital construction and operating processes.

Oberwinter therefore considers the conventional digital twin a problematic image. The idea of a single digital representation bringing together all information about a building often does not reflect reality. Different disciplines need different tools, and each system performs a different task in a building's life cycle.

Oberwinter prefers the term 'digital extended family'. Depending on the building and organisation, its members include CAD, BIM, CDE, document management, CAFM, Asset Information Management, ERP and building management systems. In hospitals, the hospital information system is another member.

The crucial question is therefore not: how do we put everything into one system? It is: how do we ensure that the different systems exchange the right information? This is precisely where Digital Information Management, or DIM, comes in.

The jungle of existing data

There is generally no shortage of data in construction and real estate. Existing buildings in particular often accumulate a mix of paper records, PDF files, CAD drawings, models, Excel spreadsheets and different software solutions over the years. Information is duplicated, contradictory or trapped in systems that do not communicate.

Oberwinter calls this the 'jungle of existing data'. More data will not solve the problem. Even terabytes of information are of little help if nobody can reliably say which file is current, which designation applies or which system contains the authoritative information.

DIM therefore does not start with data volume. Companies must first clarify which information their employees actually need, where it originates and how to deliver it reliably to where it will be used.

This is not just about software. Who creates information, who checks it and who subsequently uses it are as much a part of Digital Information Management as data models and interfaces.

'Which information do you really need?'

BIM in particular makes it technically possible to enrich a component with a wealth of information. A single digital element can contain countless technical specifications - but does subsequent operation need them all? Or would a few details suffice, such as the installer, warranty, maintenance intervals, cost centre or responsible maintenance company, together with an organised repository of technical data sheets? Oberwinter deliberately puts the key question simply: 'Which information do you really need?' Encouragingly, this concept now appears in literature and standards as 'Level of Information Need', or LOIN.

A designer works with different information from a construction company. Technical building operation requires different data from commercial facility management. The starting point should therefore be the specific use case, rather than the most extensive possible data model.

What will a user or operator want to do later? Which information will they need? Who must supply it? And when must it be available? Answering these questions early also makes it possible to define more precisely which information should be created during design and construction. The client then requests the data that will actually be needed, rather than as much as possible just in case.

The forest of standards

Anyone trying to organise building data quickly encounters the next thicket. Alongside the 'jungle of existing data', Oberwinter sees a 'forest of standards'. Standards are essential if different participants and systems are to work together. They establish common terminology and rules. At the same time, there is an enormous number of possible requirements and attributes.

Here too, more is not automatically better. Companies must decide which standards suit their specific tasks. A component needs a unique designation. Contractors need to know which information they must supply. Systems must structure data so that other applications can understand it. But that does not mean every conceivable value needs to be collected.

In other words, information needs should determine the rules, not the other way round.

The 'cave of empty promises'

Faced with a software landscape that has evolved over time, an apparently obvious thought quickly arises: why not move everything into one system? Oberwinter is sceptical. He places promises such as 'one-stop shop' or 'single source of truth' in the 'cave of empty promises'.

The different systems have valid purposes. A CAFM system supports servicing and maintenance. An ERP system handles commercial data and processes. Building management systems provide current operational data. BIM systems support design, conversions and spatial location. An Asset Information Management system brings together equipment information. Document management systems enable structured storage of content that is not organised around objects or databases.

DIM therefore does not necessarily seek to replace all these applications with a central platform. It is far more important to clearly distinguish their responsibilities and determine which information flows between them.

Each system can remain where it best fulfils its purpose. What matters is clarity about where information is maintained and how other systems can access it. The supposedly single digital twin becomes an interplay of multiple digital representations: the 'digital extended family'. Digital Information Management orchestrates this family.

We have always done it this way

Technical interfaces, however, solve only part of the problem. As soon as information flows across platforms and disciplines, organisational questions arise. Who may change data? Who may only read it? Who supplies which information? Who checks its quality? And who deals with missing or contradictory details?

For Oberwinter, this is where the 'snake pit of processes' begins. Many companies will recognise the typical reactions: 'We have always done it this way' or 'Who is supposed to do that here?'

This illustrates precisely why digitalisation is not simply a software project. When information flows differently, tasks, processes and responsibilities change too. Companies must define who enters, changes, controls, approves and checks data, and for what purpose - in other words, data governance.

This becomes particularly important in the transition from design and construction to operation. Information created by designers and construction companies will later be used by other people in other systems. This only works if everyone involved knows what they must deliver, when and to what quality.

How much BIM, and when?

Few companies can rebuild their digital landscape from scratch. Existing buildings usually come with existing software landscapes as well as data. Some applications work well and should remain; others must be replaced. There are also old CAD files, spreadsheets, documents and databases. DIM must therefore address how old and new members of the digital extended family can get along.

Vienna General Hospital (AKH) demonstrates the scale this can reach. Large datasets from an existing CAD system and many other systems had to be transferred into a new BIM-based tool landscape: more than 25,000 rooms, almost 80,000 wall elements and over 1.4 million equipment elements. A specially developed interface helped transfer the data automatically and in a controlled manner.

Digitalising existing buildings therefore rarely means discarding everything old and starting again. Companies must examine which data they can reuse, which requires cleaning and which information they will still need in future. Oberwinter also asks another very practical question: 'How much BIM, and when?' Again, this is about the right amount, not sheer volume. Not every piece of information needs to be available in the greatest possible detail at all times. Building models progressively can be effective - for example, starting with simple 'room box models' that require little effort but provide a very good basis for connecting room data across systems and modelling incrementally, such as during refurbishment or conversion.

Plandata

Never change a running system

Even when the data model, systems and processes are right, one decisive factor remains: the people working with them. New systems and workflows change everyday work. Employees take on different tasks, maintain information differently and sometimes need to learn new tools. Change is often initially met with a simple 'But it works!' or the classic 'Never change a running system!' Building acceptance, training, guidance and support must therefore be part of the process from the start. Managers must explain why workflows are changing and what the organisation aims to achieve. Oberwinter's DIM implementation model explicitly connects information strategy, organisation, technology and handling existing data with project and change management.

The goal first, then the technology

Installing new software is not enough. The initial question should therefore not be which BIM, CAFM or data management system a company wants to buy. First, the organisation must clarify what it wants to achieve: improve maintenance? Plan conversions faster? Reliably document operator obligations? Better manage energy consumption or space utilisation?

These goals produce specific use cases, which determine the information required. Only then can informed decisions be made about the data, standards, processes, roles and systems needed.

This sounds less spectacular than the vision of a single, all-knowing digital twin. Yet for Oberwinter, this is precisely the point: it is not the largest BIM model or the latest software that determines how digitally an organisation works. What matters is whether people can find and use reliable information when they need it.

The search for one perfect digital twin may therefore lead in the wrong direction. Buildings are complex, their life cycles are long and the tasks of their owners, designers, builders, users and operators vary widely.

The essential achievement is not collecting all data in one place. It is connecting the right information.

To do this, the digital extended family does not need to move into the same house. Its members need to be able to talk to one another.