T. S. Eliot, "The Rock" (1934)
Where is the wisdom we have lost in knowledge? Where is the knowledge we have lost in information?
A pyramid that is everywhere, yet rarely looked at properly
You may not have heard the name DIKW (Data, Information, Knowledge, Wisdom), but you’ve certainly seen the picture: a four-level pyramid that takes data up to wisdom. It is one of the most widely used frameworks in information science and knowledge management, and, interestingly, its name is tied to a pioneer of systems thinking, not an IT specialist.
Where did it come from?
Eliot wrote the lines above in 1934, which is the literary picture of this same idea. The formulation of the model is usually credited to Russell Ackoff, in his paper “From Data to Wisdom” (1989). It has earlier roots as well: Harlan Cleveland (1982) and Milan Zeleny (1987) had written similar pyramids. Ackoff was a theorist of operations research and an important figure in systems thinking, and I think that matters: this model was made to understand organisations and systems, not just databases.
Four levels (and one that usually gets dropped)
Ackoff knew each level by a question:
| Level | What it means | Organisational example |
|---|---|---|
| Data | Raw, uninterpreted symbols | ”120” |
| Information | Data in context; answers who, what, where, how many | ”120 orders were placed in July 2026” |
| Knowledge | Knowing “how”; a repeating pattern | ”Orders peak every Thursday” |
| Understanding | Grasping “why”; the reason for that pattern | ”Because salaries are paid at the end of the week” |
| Wisdom | Evaluative judgement; what is best to do | ”Add warehouse staff for that day, but consider their fatigue too” |
The common four-level “DIKW” version has dropped the understanding level; Ackoff’s own paper has five concepts.
Critiques that shouldn’t be ignored
The pyramid is so simple and appealing that the risk of believing it blindly is high. Some serious critiques apply:
- It isn’t linear. The path from data to wisdom isn’t a one-way staircase; often earlier knowledge or understanding decides what data is gathered in the first place (which connects to the piece on theory-first or data-first).
- The boundaries aren’t clear. The line between information and knowledge isn’t always clear in practice.
- Wisdom can’t be operationalised. Data and information can be measured, wisdom can’t; so many “turn data into insight” projects stall at the information and knowledge level while claiming insight.
What does it matter to a systems analyst?
Most failures of data-driven projects (including enterprise-system roll-outs) come from here: the organisation thinks that because it has data and reports it has insight. But having a report is far from understanding “why” and deciding well. This is the same point we made in “Automating a bad smell”: data and automated reports, if the process and decision-making behind them are faulty, don’t become insight.
DIKWdata-information-knowledge-wisdomrussell-ackoffsystems-thinkingknowledge-management