Guest Post: Not All Boxes-and-Arrows Diagrams Are the Same
Most of us are fluent in boxes and arrows.
We draw process maps.
We diagram workflows.
We show what happens first, then next, then next.
And those diagrams are useful.
But when it comes to designing high-stakes decisions — pricing, hiring, investment, policy — the diagram most teams reach for is often the wrong one.
There are actually two very different kinds of boxes-and-arrows diagrams.
One shows sequence.
The other shows causality.
Confusing them is one of the quiet reasons analytics and AI fail to influence real decisions. Surprisingly in every cohort of our Getting Started with Decision Intelligence course about 50% of the students are still confusing sequence with causality, even after 3 weeks of homework and feedback. This error is so common that we’ve given it a name, “Error 23B”.
See the Diagram Below, Top Half: A Causal Decision Diagram (CDD). Arrows mean causes. Only the first box in each chain is an action we directly control; downstream boxes represent effects in the world.
Bottom Half: A process diagram. Arrows mean NEXT STEP. Each box represents something we do.
Both are useful. However, they serve completely different purposes.
The diagram we’re used to: Process
Look at a simple process diagram for buying coffee:
Drive to store → Select coffee → Pay → Drive home.
Each box is something I do. Each arrow means: “And now I…”. Each box is an action I take that directly affects something I control—driving or walking, to a specific store of my choice, taking a bag of coffee off the shelf, tapping my credit card on the payment device.
Process diagrams are about managing activity. They’re perfect for planning a workshop, running a meeting, or documenting a workflow. They help us coordinate people and tasks.
But process diagrams only help us make planning decisions. There’s another important kind of decision that they overlook.
The diagram we actually use in our heads
Gary Klein, in Sources of Power, describes how experienced decision makers rely on mental simulation
They imagine an action.
They mentally trace forward what will happen.
They evaluate whether the outcome aligns with their goals.
That’s not a process map. That’s a causal chain.
“If I do this… then that will happen… which will lead to this… and that affects what I care about.”
That mental model is what we capture in a Causal Decision Diagram (CDD).
What a CDD really shows
In the coffee example above, suppose I’m deciding whether to buy bird-friendly, fair-trade coffee.
In the CDD my action is, “Buy bird-friendly coffee.”
That causes money to flow through supply chains.
That influences grower wages and land stewardship.
That affects environmental outcomes.
And those effects contribute to my overall social and environmental impact.
Notice something important:
When I buy the coffee, I don’t personally improve rainforest ecology.
I don’t directly raise wages.
I don’t perform the downstream actions.
My action causes effects through systems I do not directly control. That is the key distinction.
In a CDD, the arrows mean: causes. They mean “this causes…”; they do not mean: “And now I…” Once I take the action, everything downstream is an effect of mechanisms in the world.
That difference changes everything.
Why this matters for AI
Most AI and analytics tools attach themselves to process diagrams.
They optimize steps.
They automate tasks.
They speed up workflows.
But high-stakes decisions are rarely about sequence. They’re about consequences.
When a team externalizes its mental simulation into a CDD, several things happen:
Hidden assumptions become visible.
Trade-offs become explicit.
Disagreements surface early.
Data and models have a clear home.
The team becomes aligned around a common understanding
The CDD is an enabler so that
Predictive models can estimate specific causal links.
Simulation can explore alternative futures.
Optimization can search across feasible actions.
Without a causal blueprint, AI floats disconnected from decisions. With one, it becomes part of a coherent decision system.
Two diagrams. Two purposes.
Process diagrams help us manage work.
- Boxes = actions we take
- Arrows = next step
Causal Decision Diagrams help us design decisions.
- Boxes = elements in a causal chain
- Arrows = causes
- Only the first box (the action) is directly under our control
- Everything else unfolds through systems we influence but do not command.
That distinction is subtle. It is also transformative. Because once we draw the causal diagram—once we capture the team’s mental simulation—we have a shared decision design. Designs allow us to align teams and bring in assets like data, AI, and simulations to find better ways to achieve outcomes. They let us monitor and improve decisions over time. And they make decisions organizational assets that can be managed and improved to build value over time.
Nadine Malcolm
Nadine Malcolm has four decades of experience leading cross-functional, multi-national, and geographically distributed teams delivering end-to-end solutions to complex problems to customers in the US, Europe, and Asia. Before coming to Quantellia, Malcolm held management and executive positions in several Silicon Valley startups, including VP of Development at Ascent Logic Corporation which provided systems engineering tools to major aerospace companies and US government projects.
