Iceberg Model
The Iceberg Model is a systems-thinking tool that looks below a visible event through four levels: the event itself, the patterns it belongs to over time, the structures that produce those patterns, and the mental models that keep the structures in place. The four-level version is widely used in the systems-thinking community and has no single agreed originator. Working down the levels moves a group from reacting to a symptom towards the causes where a change has more effect.
Facilitation script
- 1
Introduce the four levels with one short everyday example, then agree the wording of the event at the tip.
8 min - 2
Patterns: silent note-writing on what has happened over time, then post and sort chronologically. Ask for evidence where possible.
12 min - 3
Structures: in pairs, list what in processes, rules, incentives and resources produces these patterns. Post and cluster.
15 min - 4
Mental models: whole group, ask which beliefs keep each cluster of structures in place. Write them as full sentences.
12 min - 5
Read the iceberg bottom-up, test the links between levels and mark unverified entries.
8 min - 6
Choose where to intervene: compare the usual event-level reaction with one or two interventions at structure or belief level and pick what to pursue.
10 min - 7
Assign owners for actions and for the open research questions.
5 min
Tips
Spend real time on the patterns level before going deeper; groups that skip it tend to explain a single incident and end up with opinions instead of structures.
When someone names a person as the cause, ask what in the setup would lead any reasonable person to act that way.
Phrase mental models as sentences the organisation might actually say, such as 'a late delivery is worse than a flawed one', so they can be examined.
Common pitfalls
Filling the structure level with names of people or departments, which turns the analysis into blame and leaves the actual setup untouched
Treating guesses about mental models as findings; untested beliefs attributed to others can harden conflict if they are reported as fact
Jumping from the event straight to mental models, so the result is a general complaint about culture with no link to observable patterns
Ending with a completed poster and only event-level actions, which repeats the reactive fix the exercise was meant to move beyond
Starting with an event that is too broad, such as 'communication is poor', so there is nothing concrete to trace downwards
Variations
Culture iceberg: a related two-level picture, usually traced to the anthropologist Edward T. Hall, separates visible behaviour from hidden values and is used in intercultural work; it is a different model from the four-level version described here. Remote: use a board with four horizontal bands and have people add notes silently level by level. For large groups, let several subgroups build an iceberg for the same event and compare the structure levels. Follow up with a causal loop diagram when the structure level reveals reinforcing loops worth mapping in detail.
Where it fits
When to use it
The same kind of incident has occurred several times and each previous fix addressed only the latest occurrence
A team is stuck blaming individuals for a problem that persists across different people
A change programme keeps stalling and you need to surface the assumptions that hold existing structures in place
A group is new to systems thinking and needs a simple frame before working with loops or stock-and-flow ideas
A retrospective has produced a long list of symptoms with no shared explanation
When not to use it
The problem is a one-off with a clear technical cause; a 5 Whys chain or a straightforward fix is quicker
An incident is still ongoing and needs containment; stabilise first and analyse afterwards
You need a detailed inventory of contributing factors by category; an Ishikawa diagram gives better coverage at that level
The people who know the history and the structures are not in the room, because the lower levels will then be speculation
Trust is too low to discuss beliefs openly; start with a less exposing format and return to the mental-model level later
Related methods
Frequently asked questions
What is Iceberg Model?▾
The Iceberg Model is a systems-thinking tool that examines a problem on four levels: the visible event, the patterns over time, the structures that create those patterns and the mental models behind the structures. The image of an iceberg expresses that most of what causes an event sits below the surface. Groups use it to find causes that a quick fix would miss.
Who created the Iceberg Model?▾
There is no single verified originator. The iceberg metaphor for culture is usually traced to Edward T. Hall, and the four-level systems version spread through the systems-thinking community associated with authors such as Peter Senge. Different sources credit different people, so it is safest to describe it as common systems-thinking practice.
What are the four levels of the Iceberg Model?▾
Events are what just happened. Patterns are the trends and repetitions that the event belongs to. Structures are the rules, processes, incentives and relationships that generate the patterns, and mental models are the beliefs and assumptions that keep those structures in place.
How is the Iceberg Model different from the 5 Whys?▾
The 5 Whys follows one causal chain from a single problem by asking why repeatedly. The Iceberg Model asks a different question at each level and looks at time trends, system structures and beliefs in turn. It suits recurring, systemic problems, while the 5 Whys is faster for a contained issue with a traceable cause.
How long does an Iceberg Model session take?▾
Plan 45 to 90 minutes for a group of up to fifteen. A well-defined event with data at hand can be worked through in under an hour. Allow the longer slot when the group is new to the model or the structure level needs discussion.
Plan your next workshop with AI
Workshop Weaver helps you combine methods like Iceberg Model into a complete, timed agenda in minutes.
Try it freeMethod descriptions on Workshop Weaver are original content written by our team, based on established facilitation practices. This method was inspired by work from Systems-thinking practice; no single verified originator.