Introduction

Over the past decade, observability has become one of the defining disciplines in cloud computing, DevOps and Site Reliability Engineering (SRE). Modern organisations rely on observability platforms to monitor distributed applications, detect anomalies and maintain the reliability of increasingly complex software systems [1][2].

At the same time, organisations operating physical assets face a different challenge. Manufacturing plants, utilities, hospitals, smart buildings, logistics providers and smart cities may have connected sensors, IoT platforms and operational dashboards, yet decision-makers still struggle to determine what is happening on the ground, who should respond and what action should be taken.

This gap is not a software problem.

It is an operational problem.

This paper argues that observability and operational visibility solve different classes of problems, and that Operational Blindness explains why organisations can achieve excellent software observability while remaining unable to perceive and manage their real-world operations.

Understanding Observability

Observability originated in control theory before becoming a core discipline in distributed software engineering. Today it describes the ability to infer the internal state of a software system from the telemetry it produces [3].

Modern observability platforms collect three primary categories of telemetry:

• Metrics
• Logs
• Traces

These enable software engineers to answer questions such as:

• Which microservice failed?
• Why is application latency increasing?
• Which deployment introduced errors?
• Where is excessive memory consumption occurring?
• Why are users experiencing failures?

The objective of observability is clear:

Maintain reliable software systems.

Solutions such as Datadog, Grafana, New Relic, Dynatrace, Elastic Observability and OpenTelemetry have become essential components of modern cloud operations [4][5][6].

The Limits of Observability

Observability provides deep visibility into software behaviour.

It does not explain whether the organisation’s physical operations are functioning correctly.

Consider the following examples.

A manufacturing plant reports:

• healthy cloud infrastructure
• zero application errors
• responsive APIs

Meanwhile,

• machine vibration has increased for weeks,
• product quality has begun to decline,
• maintenance has not been notified.

The software is healthy.

The factory is not.

Similarly,

a water utility may report:

• databases operating normally,
• applications fully available,
• network latency within expected limits,

while:

• a pump station has failed,
• water pressure is falling,
• customers are beginning to lose supply.

Again, the digital platform remains healthy.

The operational system does not.

Observability has successfully answered:

“Is the software working?”

It has not answered:

“Is the operation working?”

Operational Blindness

Operational Blindness refers to the inability of an organisation to perceive, understand and respond to the true state of its physical operations despite collecting operational data.

An organisation may already possess:

• sensors
• PLCs
• SCADA systems
• ERP platforms
• IoT platforms
• AI applications
• dashboards

Yet managers remain unable to answer operational questions such as:

• Which asset requires immediate intervention?
• Which abnormal condition deserves the highest priority?
• Who is responsible?
• What action should occur?
• Has the problem been resolved?
• What business impact will result?

The issue is rarely the absence of data.

It is the absence of operational understanding.

This condition is Operational Blindness.

Observability and Operational Visibility Operate at Different Layers

Although both disciplines rely on telemetry, their objectives differ fundamentally.

Observability	Operational Visibility
Software systems	Physical operations
Servers	Physical assets
Containers	Equipment
APIs	Facilities
Cloud infrastructure	Field operations
Application performance	Operational performance
Software reliability	Operational resilience
DevOps teams	Operations teams
Software incidents	Operational incidents
Digital health	Business outcomes

Observability answers technical questions.

Operational Visibility answers business and operational questions.

Neither replaces the other.

Why IoT Alone Does Not Eliminate Operational Blindness

Many organisations believe that deploying IoT sensors automatically creates visibility.

It does not.

Sensors generate measurements.

Dashboards display measurements.

Neither explains:

• whether the situation is normal,
• whether intervention is necessary,
• who should respond,
• what should happen next.

Operational Visibility requires an additional layer that converts operational data into operational understanding.

This progression can be represented as:

Without this layer, organisations often collect more operational data while remaining operationally blind.

Why AI Needs Operational Visibility

Artificial Intelligence is frequently presented as the next stage of digital transformation.

However, AI depends upon trusted operational data and meaningful operational context.

Without Operational Visibility, AI analyses incomplete, delayed or poorly contextualised information.

The progression therefore becomes:

Operational Visibility provides the operational context upon which AI can generate decisions that are relevant, timely and actionable.

Complementary Rather Than Competitive

Operational Visibility should not be viewed as an alternative to observability.

The two disciplines address different parts of an organisation.

Observability ensures software platforms remain reliable.

Operational Visibility ensures physical operations remain visible, coordinated and accountable.

Modern cyber-physical organisations require both.

Software failures affect applications.

Operational failures affect customers, safety, revenue and public services.

One protects digital infrastructure.

The other protects operational performance.

Conclusion

Observability has become an essential capability for managing modern software systems. It enables engineers to diagnose failures, maintain application reliability and understand increasingly distributed computing environments.

Operational Blindness addresses a different challenge. It explains why organisations equipped with connected devices, cloud platforms and AI can still fail to recognise operational problems before they become costly incidents.

As organisations continue to digitise factories, utilities, healthcare systems, transport networks and smart cities, success will depend upon combining software observability with operational visibility.

Observability answers the question:

“Is the software healthy?”

Operational Visibility answers the more strategic question:

“Is the organisation operating effectively?”

Only when both questions can be answered confidently can an organisation claim to have genuine operational awareness.

References

[1] OpenTelemetry. Observability. https://opentelemetry.io/docs/concepts/observability/

[2] Google Cloud. Observability Overview. https://cloud.google.com/observability/docs

[3] Rudolf E. Kalman. On the General Theory of Control Systems. Proceedings of the First IFAC Congress, 1960.

[4] Grafana Labs. What is Observability? https://grafana.com/docs/grafana/latest/explore/simplified-exploration/observability/

[5] Datadog. What is Observability? https://www.datadoghq.com/knowledge-center/observability/

[6] New Relic. What is Observability? https://newrelic.com/observability

[7] OpenTelemetry. Signals: Metrics, Logs and Traces. https://opentelemetry.io/docs/concepts/signals/

[8] Mazlan Abbas. The Canonical Definition of Operational Blindness. IoT World, 2026. (Author’s definition paper.)

About the Author

Dr. Mazlan Abbas is the originator of the Operational Blindness concept and the co-founder and Chief Executive Officer of FAVORIOT Sdn Bhd.

Drawing on more than three decades of experience in telecommunications, research, academia, IoT, and technology leadership, he developed Operational Blindness to describe a recurring organisational condition: critical operational reality remains unseen, misunderstood, delayed, or disconnected from action, even when digital systems and data are already present.

His work proposes that the value of IoT and AI should not be measured by the number of devices connected, dashboards created, or data points collected. It should be measured by whether an organisation can see what is happening, understand its significance, make a timely decision, and act before the decision window closes.

Dr. Mazlan continues to develop the theory, measurement methods, sector applications, and research agenda for Operational Blindness, Operational Visibility, and the emerging category of Operational Visibility Platforms.