Information Overload: Simon Was Right

I have, at this moment, 847 unread emails. I know this because my inbox tells me so. I stopped trying to clear that number sometime around 2023, when I realized that reading all of it would take longer than the emails took to arrive in the first place. That is not a personal failure. It is an information environment problem. And Herbert Simon diagnosed it almost 80 years ago.

Simon (1947) argued that human decision-making is bounded. We do not have unlimited cognitive capacity, unlimited time, or unlimited access to relevant information. We cannot optimize. We satisfice: we find solutions that are good enough given our constraints. He called this bounded rationality. When Simon wrote this, the information environment was modest by any standard we would recognize today. There was no email. There was no Slack. There were memos and phone calls and face-to-face meetings. He was already worried about cognitive limits in that environment.

In 2026, the problem looks nothing like 1947. The average knowledge worker, in organizations I have read about in the popular press, receives somewhere in the range of 100 to 200 emails per day, depending on their role, and that figure says nothing about Slack messages, Teams notifications, dashboard alerts, meeting invitations, and the general noise of digital work life. I am not going to cite a specific number here because the numbers vary widely by study and source and they age quickly. What I am confident about is the direction: the volume is large, it has grown steadily, and it is not going down.

Simon also introduced the concept of the scarcity of attention. He wrote that a wealth of information creates a poverty of attention. That sentence feels less like 1971 and more like an observation about my phone notifications. The insight was that information and attention are in a zero-sum relationship. More information does not automatically produce better decisions if the information exceeds the capacity to process it. At some point, adding more information starts degrading decision quality instead of improving it.

Eppler and Mengis (2004) reviewed decades of research on information overload across information science, psychology, management, and organizational behavior. Their finding was consistent across domains: beyond a certain threshold, more information makes decisions worse, not better. The relationship between information volume and decision quality is not linear. It goes up, peaks, and then declines as the amount of information passed what a decision maker can actually process. The threshold varies by person, task, and context, but the shape of the relationship holds across studies.

The irony in how we build enterprise IS is hard to miss once you see it. Every new tool gets justified partly on the grounds that it makes information more accessible. The business intelligence dashboard gives everyone visibility into the metrics that matter. The unified communications platform keeps all the messages in one place. The ERP integrates all the organizational data into a single system. These promises are real. The accessibility is real. But the tools also add to the cognitive load of every person who has to process what they surface. The dashboard I can access on my phone at midnight does not respect the limits Simon wrote about. It just moves the problem.

Gorry and Scott Morton (1971) built on Simon's intelligence-design-choice framework to classify decisions by their degree of structure. One of their core observations was that decision support systems add the most value for semistructured decisions, not for all decisions uniformly. That distinction matters. A fully structured decision has a known procedure. A fully unstructured decision barely has a defined problem. Semistructured decisions, the ones that need some judgment and some data, are where the information environment can genuinely help. But the current information environment does not sort incoming signals by their decision relevance. It dumps everything on the same channel and leaves the sorting to the person on the receiving end.

There is a commercial dimension to this that did not exist in Simon's time, and I think it changes the problem considerably. Platforms, search engines, social media feeds, and notification systems are not neutral information delivery mechanisms. They are designed to compete for attention. The business model of most digital platforms depends on it. An email provider monetizes by keeping me in the email environment. A social platform monetizes by keeping me scrolling. A news aggregator monetizes by showing me content that triggers enough response that I keep reading. These systems are not trying to help me satisfy my bounded cognitive budget. They are trying to capture as large a share of it as possible.

This is what some researchers and commentators have started calling the attention economy, the idea that human attention is a scarce resource that commercial platforms compete to harvest. Simon's bounded rationality provides the theoretical foundation for understanding why this competition matters. Attention is genuinely finite. Every minute a platform holds my attention is a minute unavailable for the task I was actually trying to do. The irony is that the same constraints Simon identified as a reason to be careful about information system design have become the target that attention-economy platforms optimize against.

For IS design, I think this creates a real obligation. The question is not only whether the system provides information people need. It is also whether the system respects the cognitive limits of the people using it. A system that surfaces everything with equal urgency is not respecting those limits. A system that sends a notification for every possible event is not respecting those limits. A system that requires a user to process thirty fields of data to answer a question that could have been answered with three is not respecting those limits. Simon was writing about the fundamental conditions of human cognition. Those conditions have not changed. The information environments we have built on top of them have.

I do not have a clean solution to my 847 emails. But I have a better understanding of why the problem is structural rather than personal.