Reasoning behind our framework

On this page, we outline the reasoning behind our framework. In particular, we explain (i) why we use a framework at all (ii) how we selected the factors (iii) our views about which factors are most important.

Why use a framework at all?

We want to compare career options in terms of how much impact they’ll let you have in the long-run. Ideally, we could precisely measure the impact of each option (perhaps in terms of how many lives you improve), and choose the best one. In practice, we’ll never be able to get a precise measurement – the impact of a career option is just far too complicated.

However, that doesn’t mean that all options are equally good – it’s obvious that some options are better than others. Rather, it means we need to prioritise what we investigate, and identify the most important dimensions to compare your options on in the limited time available. The goal of a good framework is to identify the factors that are most high-priority to investigate (essentially those which have the highest value of information).

How did we choose these factors?

What makes one factor better than another?

The importance of a factor within the framework depends on:

  • Relevance: To what extent do differences in this factor correlate with long-run impact?
  • Scale: How much do careers differ based on this factor?
  • Measurability: How is it to find out about these differences in the time available?

To take an example, whether you sit on a blue chair at work is a very easy factor to measure, but it doesn’t correlate with your impact, so it’s not important overall. Your expected earnings are somewhat harder to figure out, but matter much more for your long-run impact – earning money is useful for supporting yourself, paying for training and donating to charity. So, expected earnings might deserve a place in the framework.

When combining several factors into a framework, ideally all the factors are relatively independent. This makes it easier to prevent double counting. For instance, if you had ‘starting salary’ and ‘salary in five years’ as separate factors, there would be a danger of double counting earnings because these factors are highly correlated.

Under relevance, there’s some other rules of thumb that are useful.

  • Do decisions about this factor get ‘locked in’ or not? For instance, the organisation you work for is relatively easy to change, whereas the skills and reputation you develop will stick with you. That suggests it’s more important to get decisions about skills and reputation right first time.
  • Does this factor keep your options open? i.e. will it turn out to be useful in many future situations? We explain why this is important here.

How can we work out which factors are most important?

In line with our research principles, we look at the following sources to evidence to decide what to include:

  1. We directly estimate how well each factor typically does on relevance, scale and measurability.
  2. We consider what common sense thinks are the most important factors (we don’t think there are natural experts on this question, though we’d put more weight on the views of people who have thought hard about these issues).
  3. We consider which factors have the best ‘track record’ i.e. what have high-impact people in the past focused on?
  4. We work with the people we coach, to see which factors best capture the important aspects of their decisions.

What process did we use to create the current framework?

We aimed to choose the factors that best capture the considerations that have most often been important for the people we’ve coached, while remaining fairly independent and easy to communicate. As we explain on the main page, the framework is based on making two important distinctions: roles potential vs. personal fit, and short-run impact vs. long-run potential.

We’ve also used our background knowledge to try to directly estimate which factors are most important, searched around for relevant arguments, and tempered our findings against our impression of common-sense and their robustness. For instance, because we think it’s possible to judge that some causes are significantly more effective than others, we included ‘cause effectiveness’ under ‘role impact; although we exercise some caution because common-sense doesn’t put much weight on cause selection. We’ve outlined the reasoning behind each factor on the individual factor pages (role impact,
personal fit, career capital impact, exploration value).

Our reasoning about the relative importance of the factors is explained later on this page, along with how our views have changed over time.

We haven’t yet done much investigation of track record by examining case studies of high-impact people, and would like to do more of this in the future.

Which factor is most important?

This question can be split into two:

  1. Which factor is most important on average?
  2. How does the relative importance of the factors change at different stages of your career?

Which factor is most important on average?

For someone who hasn’t thought much about their career, we think it’s pretty clear that all the factors are important, but are unsure which is most important on average.

One significant difference between the factors, however, is that the role impact, career capital and exploration value are things that you tend to mainly investigate up-front (i.e. before you take the option) and which tend to hit diminishing returns fairly soon, whereas you can continue to learn about personal fit over many years. This gives rise to a common pattern: once you narrow down your opportunities to those that are ‘plausibly good’ based on the other three factors, then people mainly focus on learning more about personal fit. Because of this, personal fit often ends up as the most important consideration.

Which factor is most important at each stage of your career?

Briefly, we think learning more is most important at first, then it’s career capital, then it’s role impact. Personal fit is important throughout.

A sketch of our reasoning is provided on the main page.

This depends in large part on how important it is to make a difference now rather than later, which is explored on this page.

Some of the other important considerations, which we discuss on the main page, are:

  • To what extent does one factor get locked in vs. remain open to change later? Career capital seems to get more locked in than role impact.
  • Your motivation – it can be more motivating to have an immediate impact than build career capital.
  • What people know at different stages of their career – generally they don’t understand their options early on, so exploration value is important, but it becomes steadily less important as they get more experienced.

How have our views changed over time?

The first version of the framework was released in July 2013. It focused on comparing career paths in abstract, so only included role impact and career capital. When we extended the framework to cover individual career decisions, we added personal fit and exploration value.

Which factors have been included have otherwise remained fairly similar over time, though we have tweaked some of the definitions and subfactors. We’ve also experimented with different ways of grouping and naming the factors, mainly with the aim of making them easier to communicate. For instance, in the last version of the page we broke them into: immediate impact (role impact), long-run potential (career capital and exploration value) and personal fit.

Our views about which factors are most important at different stages have changed over time. In the last year:

  • We’ve come to put more emphasis on exploration value early in your career. This is mainly because we’ve become more aware of the challenges of finding a good fit, and more keen to encourage people to consider more options.
  • We’ve come to put more weight on career capital early career, because we came to put more weight on investing in yourself rather than having an immediate impact.
  • We’ve come to more weight on personal fit throughout. We thought it was important to think about your chances of success in different careers before, but we’ve now become more aware of the indirect benefits of personal fit, largely due to discussions with Holden Karnofsky.

Directions for future research

Our main uncertainties involve the relative importance of the factors at different stages of careers. Unfortunately, this question involves several big judgement calls that are difficult to investigate empirically, however, we intend to:

  • Continue testing the factors in our coaching, to see whether we’re missing something important, and to continue to get a better sense of how easy it is to learn more about each factor.
  • Do some case studies of high-impact people to try to see what factors they focused on.
  • Do more career profiles to learn more about the differences in role impact between common career paths, and to learn more about how to determine personal fit between people and different career paths.