Refiner: Using Filters to Design a Decision Model

Analyzing options and deciding between those options take place in different areas of the brain. Analysis is rational, logical, and language-based. Deciding is intuitive and driven by gut feeling. Creating a decision model is a way to blend both analysis and intuition. It pulls decision making out of the black box of intuition and into the plain view of the group.

View fullsize

1. Begin by giving the decision model a title that labels its overall purpose. For example, the purpose may be to decide initiatives for a strategic plan, or where to focus R&D resources, or prioritize a backlog of change requests.
2. Have participants list a dozen or so options, writing each one on a sticky note. Place these stickies on the left side of a whiteboard.
3. Draw a series of diamonds from left to right across the whiteboard. Explain to participants that a decision model is like a vacuum cleaner with a series of filters. The goal is to create filters that are porous enough to let the right options pass through and dense enough to catch the less-desirable options. Ask participants to write two or three options that they clearly do NOT want to make it through the model to use to test the model and place them with the other options on the left of the whiteboard.
4. Have participants brainstorm criteria for the decision model. Cluster similar criteria. For example, a group might say, “We need to make sure that what we do is valuable.” Write VALUE at the top of one of the diamonds. Ask the group how they might define value. “Customers are willing to pay for the service.” “At least half of our existing customer base will use it.” Write down those details underneath the word VALUE and in the diamond.
5. Pick up a sticky and move it to the first filter. Ask, “Is this option VALUABLE?” quickly followed by “Are customers willing to pay for it?” Watch the reaction of the group. If it’s a unanimous yes, move on to the next question. If it’s a unanimous no, leave the sticky where it got “stuck” by the filter. If there’s disagreement, ask the group to refine the criteria. Do they need to define the amount customers are willing to pay? The object here is not to pass the options through the filters, but rather to refine the filters. Test the filters by using the two or three options clearly NOT wanted.
6. Limit criteria to answering only one question at a time. Try to limit criteria to binary (yes/no) results. For example, “Will more than half of our customers use it” is better than “how many customers will use it.” Again, adjust the criteria if too many or too few options are passing through the criteria: “Will more than 75% of our customers use it?”

Iterate on the decision criteria by passing different options through the filters. At first, the goal will be to create a baseline decision model, not make decisions. Over time and after several iterations, participants will become comfortable using the decision model for its intended purpose.

One variation of this method is to give thought to the horizontal placement of filters against one large factor, such as placing “must do” criteria on the left and “nice to have” criteria on the right. This results in options being organized from left-to-right against that large factor.

When considering different projects or initiatives as options, here are some criteria to consider:

• Is the option valuable? Is someone willing to pay for it?
• Will X number or percentage of our customers buy it?
• Is the option rare? Does it not already exist somewhere in the marketplace?
• Will it be difficult for a competitor to duplicate the option?
• Are we organized in a way that can execute the option?
• Can we afford it? Do we have the resources available? (time, knowledge, money, facilities, supply chains, workforce)
• Does it help us achieve our vision, mission, or purpose? Does it align with our values? Does it align with our brand?

Participants may resist the process of designing a decision model. This is usually because the model doesn’t pass their own “gut sense” intuition, or because the model will have outcomes that are not in their favor. In either case, participants must feel safe enough and have sufficient self-awareness to bring those reactions out into the room and make them explicit, which can shine a light on new aspects of the decision model.