The problem with today’s information systems landscape is rarely the systems themselves. The real challenge is that organizations lack a common governance of the information that the systems are supposed to carry. Unclear data ownership, local definitions, Excel-based shadow systems, and fragmented authority create parallel views of the business – with increased risks, inefficient processes, and uncertain decisions as a result. This article explores why local organizational optimization often stands in the way of an organizational-wide information model and why Data Governance must become a strategic management issue. Only when responsibilities, concepts, master data, and architecture are aligned can a real ”Single Source of Truth” become possible.
”Fucking shitty system”
The title is taken from a book with the same name, written by Jonas Söderström. For a long time, many organizations have invested in large systems such as business management systems (ERP, Enterprise Resource Planning) such as SAP to create an integrated information platform for finance, logistics, purchasing, production, and sales. Despite this, much of the actual analysis, reporting, and decision support takes place outside the ERP system, often in locally developed Excel sheets and proprietary reporting solutions.
The consequence is that the organization lacks a common and coordinated information model. The same business event can be described and interpreted in different ways depending on which system or report is used. A common example is the delivery date, where the definition can differ between order management systems, transport booking systems, and ERP systems. This leads to different departments presenting different versions of the same facts.
The problem is exacerbated by the fact that users have to work in many different information systems with separate user accounts, authentication solutions, and access models. Even when the systems technically have high availability, operational problems arise due to login difficulties, access conflicts, and complex workflows. For the employee, the systems are perceived as inaccessible even though the servers are functioning.
From a business perspective, this leads to:
- Increased administrative burden
- Lower productivity
- Higher operational risk
- Impaired data quality
- Increased personal dependence
- Weaker decision-making basis
- Stress and frustration among users
The central problem is therefore not that the systems are down, but that the organization lacks a cohesive information architecture and clear governance of data as a strategic asset.
Historical overview
Historically, ERP systems were built to be the organization’s ”single source of truth”. When SAP and other ERP platforms were introduced, the goal was to consolidate processes and information flows into a common system.
In practice, however, the needs of businesses have evolved faster than the ERP systems’ ability to adapt. New business models, new regulatory requirements, digitalization, cloud services, and specialized business systems have led organizations to gradually build an ecosystem of complementary solutions.
The result has often been:
- ERP system for transactions
- BI system for reporting
- CRM system for customer data
- Logistics systems for transport
- Portal solutions for suppliers
- Excel for analysis and operational control
Each introduction has been rational from a local perspective but has at the same time increased the overall complexity.
Many organizations have therefore moved from an ERP-centric approach to a situation where data is scattered across many systems. As data volumes grow, questions arise about:
- Data definitions
- Data ownership
- Responsibility for master data
- Information security
- Data quality
It is common for management to only discover the problem when different reports show different numbers for the same KPI.
CoS's view on the rotor problem
Reason 1: Unclear data ownership
Perhaps the most important reason is that no one owns the information from a business perspective. IT owns the systems. The business creates the information. Analysts process the information. But who owns the data itself?
Examples:
- Who owns the customer number?
- Who owns the delivery date?
- Who owns product information?
Without clear data owners, different definitions arise.
Reason 2: Lack of Master Data Management
The organization often lacks a central model for:
- Customer data
- Supplier data
- Product data
- Organizational data
Different systems then develop their own versions of the same information.
Chapter 3: Excel as a shadow system
When users do not get the analysis they need from ERP, they create their own spreadsheets.
Advantages:
- Quickly
- Flexible
Disadvantages:
- Version problems
- Manual errors
- No traceability
- Addiction to people
Excel then becomes an informal information system.
Reason 4: Fragmented identity management
Many logins mean:
- Multiple user accounts
- Multiple passwords
- Several authority structures
The user's experience of availability is impaired despite high technical operational availability.
Reason 5: Inadequate information architecture
The systems have been introduced gradually without a unified control system.
Consequences:
- Data is duplicated
- Integrations are growing unchecked
- KPI definitions differ
- Decision-making becomes inconsistent
4. Who owns the data and information?
This is a key Chief of Staff issue.
Internally
Recommended model:
| Information | Business owners |
| Customer data | Sales Manager |
| Supplier data | Purchasing Manager |
| Product data | Product owner |
| Logistics data | Supply Chain Director |
| Financial data | CFO |
| Personal data | Human Resources Director |
The IT department should be the system owner and infrastructure owner, but not the information owner or the workflow owner.
Externally
Information acquisition becomes more complex when information comes from external actors:
Examples:
- The Tax Agency
- Banks
- Carriers
- Suppliers
- Customers
- Authorities
The organization still is responsible for the quality of its own processes even though the basic information comes from an external source.
Name change and connection to the population register
An illustrative example of information management is name changes after marriage. In Sweden, the name change is registered with the Tax Agency through the population register. If the company’s HR system is integrated with the population register data, the name change can be automatically updated in certain systems. The problem is that large organizations rarely have a single identity. The person can appear in:
- HR system
- Active Directory/Microsoft Entra ID
- SAP
- Wage system
- Eligibility system
- Time reporting system
- CRM
- Collaboration platforms
If identity management is not well integrated, a surname change can create:
- Duplicate accounts
- Incorrect permissions
- Failed integrations
- Limit access to one or more systems during the week
- Problems with signing and authentication
This illustrates an important principle:
An information object (a person) must have a common identity across the information landscape.
Rotor cause 6: The conflict between business-based and infological implementation
One of the most strategic reasons for fragmented information systems is the choice between a business-based and an infological implementation model.
Business-based implementation
In a business-based implementation of an ERP system, each organizational unit is allowed to implement the system based on its own needs, processes, and local priorities. Each business area, division, or country can then create its own:
- customer number
- SKU
- supplier number
- organizational structures
- Product classifications
- report models
From a project perspective, this appears attractive because:
- the implementation goes faster
- Local resistance is decreasing
- existing working methods can be retained
- the business experiences a high degree of autonomy
In the short term, the project cost and resistance to change are reduced. In the longer term, however, a significant information problem arises. Although the entire organization uses the same ERP platform, such as SAP, in practice there are in fact several separate systems. What is called ”one SAP system” often consists of several independent installations or logical environments with different data models, different code customizations, and different definitions of business concepts. The consequence is that the organization achieves only local optimization while the possibility of group-wide management is impaired.
Infological implementation
In an infological implementation, the organization's information is considered a shared strategic asset. The implementation then proceeds from questions such as:
- What is a customer?
- What is a product?
- What is a supplier?
- How is the delivery date defined?
- How is turnover defined?
- How is inventory status defined?
The goal is to establish common master data for the entire organization. This means:
- a common customer register
- a common article register
- a common supplier register
- Common KPI definitions
- Common processes
The entire organization then works within the same information structure regardless of organizational affiliation. The benefits are significant:
- a real ”Single Source of Truth”
- higher data quality
- simpler analysis
- better AI and automation capabilities
- Improved governance and transparency
- simpler compliance with regulatory requirements
The organizational dilemma
The problem is that the infological model requires something that many organizations find difficult to achieve, which is organisational-wide agreements. A common data model means that local businesses have to compromise.
For example, it can:
- The sales organization wants to define customers in a certain way
- the logistics organization on another
- the financial organization on a third
Therefore, infological integration is more about governance and leadership than about technology. The biggest cost is usually not system development but organizational coordination.
Accumulated technical debt over time
When several business-based implementations are allowed to be developed over many years, a growing technical and informational debt arises. Different parts of the organization will gradually:
- upgrade at different times
- develop your own add-ons
- customize data database schemas
- introduce their own integrations
- change the process logic
After five, ten or maybe fifteen years, two SAP installations within the same group can have as many differences as two completely different ERP systems.
The organization then receives:
- several pieces of information truth
- several integration platforms
- several development models
- several version chains
The technical platform is shared by the name but not by the functionality.
Transformation cost
When the management later decides to create a truly organization-wide information system, one of the most expensive types of digital transformation arises. The transition from several business-based implementations to a common information system requires normally:
- Harmonization of master data
- consolidation of processes
- Migration of historical data
- rebuilding of integrations
- changed governance models
- Comprehensive change management
In large international organizations, such transformations often amount to hundreds of millions of euros and in some cases billions of euros. The actual cost is not the technology but the fact that the organization must agree on a common description of reality.
Summary
The central business problem is not a lack of functionality or low availability in ERP systems like SAP, but the lack of a cohesive management of information as a strategic asset. When data ownership is unclear, common definitions are lacking, and Excel is used as a shadow system, parallel versions of the same business information emerge. Fragmented identity and authority management, as well as weak information architecture and master data management, exacerbate the problem. The consequence is higher operational risk, inefficient processes, and decisions based on different perceptions of reality. The underlying cause is often the tension between local business optimization and organizational-wide information management. If ERP is implemented based on the needs of separate businesses rather than a common information model, local solutions and proprietary truths are built into the structure. The longer this is allowed to continue, the greater the complexity and future transformation cost of creating a truly organization-wide information system. Therefore, the priority for a Chief of Staff should be to establish a clear Data Governance model. Every critical amount of information needs a designated business owner, common concepts and rules for how information is created, quality assured, changed, shared and used. This must be combined with common master data, a cohesive information architecture and consistent governance of access. The crucial strategic question is whether the organization should operate as a federation of businesses with their own truths or as a company with a common information model. The answer determines whether a ”Single Source of Truth” can become a realistic foundation for digital governance, effective collaboration and reliable decision-making – or remain a vision that cannot be implemented.
