Strategy & Criteria Decision Matrix — A Framework for Decision Making
Update: This post catalyzed an interesting conversation on Hacker News.
I attended a talk by Evan Moore a while ago, where he described a framework the early team at Doordash used to make decisions and prioritize growth strategies. I was impressed by the systematic way of listing strategies and assigning them weights based on possible impact + predicted effort it would take to execute those strategies.
Inspired by that framework, I came up with a twist that generalizes the same idea to reason more broadly about decision making. The framework involves decomposing decision making into thinking about criteria and performing careful weight assignment to aggregate the preferences of all stakeholders involved.
I recently tried this idea out with a group of friends to prioritize different strategies that we should act on as a company. We all walked out of the meeting with a richer understanding of the tradeoff space between different strategies, and with a feeling of satisfaction with the process by which we arrived at the final strategy.
I thought it’d be helpful to share this framework more broadly —
Caveats
 This approach might seem overly quantitative, but the exercise of assigning numerical weights forces stakeholders to surface internal preferences that otherwise would’ve remained hidden.
 Make sure you invest enough time choosing the proper criteria to evaluate strategies!
Note
An algorithm is provided at the end of this post. I thought it’d be simpler to understand the process through an illustrated walkthrough. Feel free to skip ahead if you wish to read a more concise representation.
Walkthrough: Climate Action Strategies
Context
Assume you were to evaluate different strategies for climate change action. It can be overwhelming to decide which one to prioritize with so many proposed approaches ^{1}.
Walkthrough
 Concisely state each strategy:
Let’s pick three strategies that could help us make effective climate change action: carbon accounting, carbon capture, and fusion power ^{2}.
 Generate a list of criteria to evaluate the strategy against ^{3}:
Here are six criteria that you might want to evaluate climate change action strategies against: technical risk, impact  short term, impact  long term, capital requirements, ease of organizational support, and ease of scaling.
You should rewrite the criteria such that a higher criteria score (described later) would be good. So, in this case, we should rewrite “capital requirements” to “capital ease” and “technical risk” to “technical ease”.
 Assign an importance score to each criterion ^{4}:
The importance score is how important a given criterion is. This step is insightful since it surfaces what different stakeholders prioritize when choosing a given strategy. For example, one stakeholder might not care too much about the shortterm impact that a climate action strategy has but cares deeply about the longterm impact.
Taking the average weight score across all participants is suggested.
 Evaluate each strategy against the criteria, and fill out a criteria score for how well the strategy aligns with the corresponding criteria:
This step will typically engender intense debate (by design)! Assigning numerical weights will force stakeholders to surface internal preferences. From experience, filling out the elements of this matrix can take a while.
At this stage, when assigning numerical values to cells, it’s helpful to create bulleted notes for each stakeholder’s view. In my experience, this is a huge driver of this strategy’s value since it forces the conversation of: “for a given criteria, and a given strategy, what does a given person think?”. For example, when debating the ease of organizational support (e.g. from government, nonprofits, society) for fusion power, some participants might think that support would be tough to garner (from historical precedent). Still, some participants might think that support be simple to garner (due to pending climate change legislation). Thus, the process of filling out cells is effectively one of aggregating all participants' views.
Taking the average criteria score across all participants is recommended ^{5}.
 Do a weighted average (importance score * criteria score) for each strategy.
Straightforward. Just multiply the corresponding numbers from the weights and strategy columns.
 Select the topranked strategy.
According to the numbers I filled out, it seems that Carbon Accounting would be the best strategy to go after. This outcome was interesting considering that, intuitively, working on Fusion Power seems like a better (and more exciting) idea.
At this stage, one path is to reconcile why the topranked strategy differs from what I intuitively feel should be the topranked strategy. I may think of more criteria to evaluate a strategy against or I may have to recalibrate the weights assigned to different criteria ^{6}.
Algorithm
 Concisely state each strategy (matrix columns)
 Generate a list of criteria to evaluate the strategy against. (matrix rows)
 Assign an importance score to each criterion.
 Evaluate each strategy against the criteria, and fill out a criteria score for how well the strategy aligns with the corresponding criteria.
 Do a weighted average (importance score * criteria score) for each strategy.
 Select the topranked strategy.
Key Takeaways
Hopefully, you will walk away from this exercise with a richer understanding of the tradeoff space between different strategies. Furthermore, you will likely gain a more nuanced understanding of the internal preferences of all stakeholders involved in making the decision.
This exercise also helps avoid a “blind men and the elephant”, like situation where the discussion doesn’t come to a resolution, because you’re actually debating about different criteria! Additionally, this method helps all key stakeholders have their voices heard and see why people are making choices, which is great for team morale and for increasing the likelihood of making good decisions.
You might even arrive at the same decision you would’ve made without engaging in this exercise. However, it’s almost guaranteed that all participants will walk away gaining a deeper appreciation of the outcome.
Thanks to Max Novendstern for reading an early draft of this post and contributing many key observations.
Footnotes

In reality, we’ll need to allocate resources to multiple avenues in parallel to make substantial strides towards solving climate change, but let’s ignore this for simplicity. ↩︎

Neat projects in this space I’m excited by are Watershed (carbon accounting), Prometheus (carbon capture), and Helion Energy (fusion power). ↩︎

Spend plenty of time on this step, since it deeply influences which strategy is chosen! These criteria should ideally be mutually exclusive and not have overlapping themes. One pitfall to guard against is having overlapping criteria, since this results in overcounting the influence of certain themes. ↩︎

Rewrite the criteria so that higher weight is good. For example, if the criteria are “technical risk” rewrite this as technical ease. ↩︎

As an extension, one could try experimenting with different weighting schemes at this step too, instead of a simple average. ↩︎

I proposed this as I was a solo participant in this exercise. If there are more stakeholders, this would be expensive to do. ↩︎