The Looking Glass: Chains of Assumptions
How to avoid the false veneer of perfection
Dear readers, instead of my usual short snippets on a theme, I’m sharing an in-progress essay on a topic near and dear to my heart. A short summary and set of implications are behind the subscriber paywall. As always, thank you for your reading and support. Sending you much love and warmth during this holiday season.
Update: Part 2 is available to read here.
One of our team members, Nishant, asked me a question in a 1:1 recently: “How can I have impact?”
Now, this is the kind of question managers are asked so often it’s usually stored in the caches of our brain. So I opened my mouth to rattle off a usual answer, something like Solve customer problems through shipping great things while being a team player.
But this time, I stopped and really thought about the question.
What is impact?
The dictionary defines it as: the effect of a person, thing, or action, on another.
There are two key words in this definition:
- The first is effect — we must be able to see (or measure) some outcome.
- The second is another — we must define some target we are looking to impact.
We can’t talk about impact unless we talk about both of these things.
So in the context of my conversation with Nishant, what is the target we have in mind?
Because we are working together at a for-profit company operating in a capitalist economy, our answer is the same as any other such company — grow profits in the long run.
So the truest response to How can I have impact? is this:
Do things that contribute to profit growth in the long run.
Yes, this is the first-principles answer, the kind of answer I can 100% confidently stand behind!
Alas, it isn’t terribly useful.
I can imagine Nishant thinking Wtf do I do with that? It gives him no guidance on his day-to-day.
But Nishant is a sharp cookie, so he’ll probably immediately launch into the next volley of questions:
- What things contribute to profit growth?
- What is “the long run?
These are excellent questions! But I am now faced with a problem: whatever answers I give will no longer be guaranteed to be true.
Whatever answers I give will start relying on my chains of assumptions.
A useful answer to “How can I have impact?” is one that is specific and actionable.
Something like “Do Task X for Project Y in Z days” fits the bill.
The problem is, how do we know doing X will actually be impactful?
We don’t, at least not with 100% certainty.
What we have is a chain of assumptions that might look something like this:
- X should be done at a high level of quality (better than competitor Q).
- Z days is an ambitious yet realistic target for doing X well.
- Doing X is a significant contributor to Project Y succeeding.
- Project Y succeeding will have immense value for Customers A, B, and C.
- Customers A, B and C are the ideal customers for us to target right now.
- These customers are ideal because we think they represent many other companies like them in the future.
- These customers are ideal because we anticipate being able to charge them N amount in the future.
- The time frame of “the future” that matters to our planning at this phase is the next 1 year.
- Focusing on delivering customer value is the most direct way for us to charge them N amount.
- Charging N amount strikes the best balance between retaining and growing customers in the next year while keeping us solvent.
- Focusing on retaining and growing customers while being solvent in the next year is the best short-term strategy for long-term profitability.
Whew! I’ve only listed 11 assumptions in this chain but you can probably see that examining each question, like zooming into a fractal, only begets more questions.
For example: X should be done at a high level of quality (better than competitor Q) can be broken into:
- Competitor Q is the ideal benchmark
- High quality is important for Project Y succeeding
- Fast performance is an important aspect of high quality
- Aesthetically-pleasing is an important aspect of high quality
- Aesthetically-pleasing means not having too many choices present on the page
… and so forth
Overwhelming isn’t it?
This is made even more so when you realize a) wow, it’s hard to even put into words what I am assuming and why I am assuming it and b) there is no end to the chain of assumptions!
Indeed, just like a fractal, we can go infinitely deep.
For example, if we assume we should code Project Y in Python, that will be based off of assumptions that Python is reliable, which are based on assumptions about compilers, how computing works, how circuits work, how the laws of physics work, and so forth.
When we take a step back, it’s humbling how many of our assumptions need to be correct for “Do X for Project Y in Z days” to ultimately result in “Contribute to profit growth in the long run.”
If any assumption is substantively wrong enough, it might be the defective link that breaks the entire chain, leading us to fail in our objective.
This is why failure is the most likely end result of most new companies or projects. If it were easy, all of us would be swimming in profits by now ;)
At this point, I hope I’ve convinced you that a) whatever assertions we make are based on a long chain of assumptions and b) it’s probably the case that there’s mistaken assumptions somewhere in that chain.
The adage Think from first principles is meant to counteract the above. Don’t just accept my (or anyone’s!) word for it that Do X for Project Y in Z days will lead to Profit growth in the long run!
Think from first principles reminds you — as a smart, independent thinker — to be skeptical, to not take anyone else’s assertions on its face value.
Instead, start from known truths (aka “first principles”) and build the assumption chain link by link. For each assumption, ask yourself “Do I believe this? What explanation convinces me to believe in this?” The more convincing the explanation, the stronger the link in your chain.
Layer by layer, you work yourself up to conclusions you can get behind, keeping in mind the strength of the entire chain.
Update: Read Part 2 here
Imagining the work that I do every day out as identifying and validating my Chain of Assumptions has changed the way I approach my work.
There as many implications to this metaphor (and more I am still uncovering). A few of the biggest ones:
