Problem Formulation

Step 3: Add Information Flow and Iteration

Once participants have identified the major decisions, they add arrows showing how information moves between them.

Ask:

What information is passed from one decision to the next?

Use arrows to show flow.

Add loops when users need to revisit criteria, regenerate options, or refine selections.
When you think that users might run out of options after doing an activate, then connect the activate decision to a create decision.
When users might want to adjust the threshold of an activate decision add a self loop for the activate decision.

This step helps participants see that decision-support tools are not only about showing data. They also need to support transitions between decision stages.

Home-buying Example

The decisions mapped in Step 2 can be turned into a workflow. The workflow is not just a linear sequence; it also includes loops when the option set is too small or when the buyer’s criteria need to be revised.

Below are some example workflows:

Simple workflow

homebuying1

In the home buying example we could have create homes, where we generate 3 options A, B and C.
Then activate, where we compare each home against the price threshold, in this case home B is filtered out.
Then choose the best home, where maybe even though both homes are affordable, C is cheaper than B but also closer to work.

Iteration

homebuying2

We can also represent how someone would iterate on a decision.
For example, when activating by price, we might start by filtering out homes priced over $1 million.
If that leaves us with no available options to choose from, we’d go back and adjust the threshold — say, increasing it to $2 million, until some homes meet the criteria.
On the diagram, this looping back is shown with a backward arrow, representing how users revisit and modify decisions

Repeated Decision Types

homebuying3

This framework isn’t just about the CREATE → ACTIVATE → CHOOSE pattern.
In practice, these tasks can appear in any order or combination
For example, we can use repeated tasks to represent different thresholds we are using in the activate decision.
One for price and another one for the number of bedrooms.

Different Decision Goals

homebuying4

We can also change the goal of the decision process
In this case the goal is to CREATE.
For instance, when a real estate agent prepares a report of homes for their client, they’re generating new information rather than selecting a single option.

Collaborative Decision-Making

homebuying5

Other configurations can also represent decision-making processes that are collaborative, involving multiple people.
For example, when a family is searching for a home together
Each member goes through their decision process and shortlists their favorite options, and then they come together to compare and decide on one home to buy.