Using a Decision Architecture
Here at the Uncertainty Project, we’ve talked about the importance of “deciding how to decide”. Crafting a decision architecture is how you introduce a decision-centric approach to strategic leadership, and the use of data.
But what does it “feel” like to apply elements of a good decision architecture? What would be different from the current state or current practice around decision making? We will walk through an example, covering a “week in the life” of making a decision.
But first, let’s dig deeper into how a decision architecture relates to the concept of choice architecture.
While most discussions of choice architecture focus on consumer behaviors (like buying something at the store, or selecting a health plan), the ideas can be applied to any decision context, including leadership decision making.
This is tricky, since we associate the word “choice” with the options of a specific decision. But remember that we also make choices on how to manage information (and collaborate with others) to attack a decision-to-make. This is the realm of decision architecture.
Another way to think of decision architecture is to see it as the general choice architecture that steers the creation of specific choice architectures for each decision.
Choice Architect: Someone who has the responsibility for organizing the context in which people make decisions. (“Nudge”, Richard Thaler and Cass Sunstein)
Once you understand the concepts, you start to see choice architectures (created by others) everywhere, influencing how you decide. For the most part, we are always living in other people’s choice architectures!
When you are helping make a decision in an organization, you are actually living in two distinct choice architectures:
- Decision Architecture: The choice architecture that offers you choices for decision support, through the explicit (or implicit) decision architecture. [Designed by the organization’s decision architect(s)]
- Decision-specific Options Assessment: The choice architecture that outlines the options/choices for the specific decision-to-be-made, constructed by the decision makers. [Designed by the owner(s) of the specific decision]
In Richard Thaler and Cass Sunstein’s book “Nudge”, they outlined eight tools or elements to guide choice architects:
- Understanding defaults - choices that people must take active steps to avoid, like the status quo.
- Expecting errors - mistake-proofing the use of the decision architecture, where possible
- Providing feedback - short feedback loops improve decision making
- Understanding mappings - making information about various choices, and how they relate to outcomes, more comprehensible
- Structuring complex choices - as number of options grows, ability to influence the decision can grow, so presentation matters more.
- Understanding incentives - asking “Who chooses?” “Who uses?” “Who pays?” “Who profits?” for a specific choice architecture
- Curation - customize or specialize the choices, to match the needs or desires of the targeted choosers
- Make it fun - go byond “making it easy” to choose, and make the desired activity enjoyable
These ideas are relevant for both the construction of a decision architecture, and for the execution of a specific decision. In this post, we will focus on the former.
A Hypothetical Decision Architecture
Next let’s look at a hypothetical decision architecture, and walk through a “week in the life” of a single, important decision for an organization. We’ll stress the behaviors, information structures, and collaboration models, but avoid the specifics of the decision itself.
Let’s assume this is a difficult, strategic decision - something that will drive the allocation of significant resources in the organization. It’s worthy of a collaborative, thoughtful approach to making a choice.
This decision architecture diagram shows the evolution of information, over time, via collaboration, punctuated with a decision:
Let’s walk through it, from left to right.
- Framing - The originator of the decision starts to capture the challenge at hand, and the characteristic of the choice to be made. This framing will confirm that this is a crucial, consequential decision, that deserves a diverse set of perspectives, and needs some time to arrive at a common understanding of the problem first.
- Timeframe - This hypothetical decision architecture uses a concept of a Decision Campaign to set expectations on how long it will take to arrive at a choice. In a hybrid, distributed setting, it can leverage asynchronous communications, and a couple meetings, to balance between “enough time to think about it” and “promptly making a call”. In our example, we create a campaign lasting a week.
- Roles - The originator might start to assign participants into specific roles, as defined by their decision making practice. Who in the organization has the decision authority or decision rights to this choice? Make them the “decider” (green). Who can bring diverse, relevant perspectives that improve the quality of the decision? Invite them in and make them “Contributors” (various colors). Who else has a stake in this decision? Make them “Stakeholders” (purple). Who will ultimately implement this decision (and is not a contributor)? Add them to the “Informed” list (various colors).
- Leading with Questions - To open the campaign, the decider asynchronously distributes the framing of the challenge and the nature of the decision to be made. They solicit questions from the contributors to capture the “known unknowns” and size up the gaps in understanding across the campaign group. They provide time to let the group offer answers (some contributors are experts), and build a decision-specific FAQ, for reference.
- Dialog - The decider schedules a meeting where the desired outcome is to gain a shared understanding of the “challenge space”. No brainstorming of options or ideas, just active listening to synthesize the diverse perspectives. Have all the questions been answered? Probably not. But there should be a common outlook on the completeness of the information available to the group. The group can also (collectively) begin to make predictive judgments on their sense-making around the decision.
- Criteria - The decider applies what was learned in the dialog and defines what is most important in a choice, with a set of criteria that will be used to evaluation the ideas and options that emerge. It is critical that these are defined and distributed to the group prior to idea generation.
- Divergent Ideas - The decider (or a skilled facilitator) schedules either an asynchronous, time-bound idea generation activity, or a synchronous meeting. This is a tightly coordinated, highly structured exercise. It seeks a broad range of ideas, and relies on the facilitation or collaboration structure to minimize the impact of individual bias on the inputs.
- Convergent Opportunities - The group works to narrow the ideas to a set of the most promising opportunities, for further assessment. For each opportunity, the group works to identify the underlying assumptions that would have to be made. They ask, “what would have to be true for this opportunity to be successful?” The group knows that for strategic decisions, these opportunities should outline different paths, not present overly-specific alternative plans. Additionally, the decision architecture mandates that the status quo - or “do nothing” option - is included as a competing “opportunity”, to minimize the status quo bias.
- Deciding - The decider leverages the criteria, opportunity assumptions, the FAQ, and all dialog and discussions along the way, to make a choice. They give themselves a day to digest it all and get some distance from the emotion of the dialog and evaluations. The assumptions associated with the preferred choice are evaluated as risks, and documented. The decider knows that a decision should offer a guiding policy to the strategic challenge, and can explain why some options were NOT chosen as well. They also know that strategic decisions are often backed (and clarified) with changes investment levels.
- Commitment - The decider presents the choice back to the participants in the decision campaign. In many cases, the decider will seek a commitment to execute the decision, before moving on with wider communications. In their decision architecture, they have made it clear that they do not need consensus, but want to voice the dissenting opinions as some might “disagree and commit”.
- Communication - The decider distributes a summary of the decision made, the rationale behind it, and the next steps needed to trigger execution. This is made available to the stakeholders and those that need to be informed (at a minimum), and perhaps opened up to general consumption, depending on sensitivity.
Evaluating a Decision Architecture, as a Choice Architecture
In this hypothetical example of a decision architecture, we can explore several of the elements of choice architecture introduced by Thaler and Sunstein.
- It has sensible defaults for collecting framing information, setting a deadline, leading with questions, and setting clear roles.
- It expects multiple contributors, to improve predictive judgments when answering questions, and minimizing errors during this sense-making activity.
- It is highly collaborative, so questions, answers, ideas, and rationale can receive prompt feedback.
- It demands that ideas are evaluated against criteria that shows the mapping to desired outcomes or definitions of success.
- It should use the framing information to make it easy for the decision maker to choose how to balance speed and deliberation with a simple presentation of techniques.
- It aims to align incentives by making the Decider [who chooses] also (1) accountable for the result [who profits], (2) responsible for the follow-up actions [who uses], (3) and have budgetary authority over the execution [who pays]
- It has been curated from the rich set of decision making techniques available to match the needs of the organization, and their skills and abilities.
- It has emphasized collaboration and truth-seeking, to keep leaders from stressing about making “perfect” decisions with incomplete information. This focus on inquiry and shared learning is less pressure-packed and makes the process more fun.
What does your current state of practice for decision making look like?
Can you draw a similar decision architecture diagram to show how your organization structures collaboration, to evolve information over time, and punctuate the evolution with a decision?
Share in the comments below and show us your diagrams!