High Quality Writing
Writing well is important for at least three reasons: the process of writing, and writing well, helps you think more clearly and develop better, more complete ideas; clear and crisp written communication is the most efficient way to transmit these ideas to someone else, especially when the ideas are complex; finally, leaving behind a written document memorializes a line of thinking that will be available for future people (that future person being you, perhaps) to inspect.
Writing is often the best way to think. People are limited by small working memories and overactive brains that skip around from thought to thought. As a consequence, muddled, incomplete, self-contradicting ideas can easily survive in your own head. They can also easily survive in conversation for the same reasons. But they cannot survive careful writing and reading.
As Francois Chollet has said:
In particular live thoughts have limited recursion depth, and only have access to limited working memory; written thoughts have unlimited recursion depth, and paper has unlimited memory.
— François Chollet (@fchollet) April 23, 2018
It’s important for us, in particular, to write well, since we do so much of it: internal memos (like this one); JIRA descriptions, summaries, and comments; Slack back-and-forths; proposals and deliverables to clients; etc. And the ideas that we’re communicating are complex; think, for example, about how much is going on when trying to explain a list experiment or a conjoint output to a client, or to someone new at Gradient.
Since we do so much writing and so much of it is internal, it may be tempting to try and punt writing things down and just have a video chat and discuss the issue. The problems with this, however, are that:
- It’s often not clear what the scope of the conversation is (one person may be thinking the conversation is about how to do X, the other that it’s about whether we do X), so in order to have a productive conversation there is writing that must be done to prepare the conversation
- It leaves nothing for a third party to inspect and understand
- Without something in writing, two people can leave a conversation with different understandings of what was discussed and agreed-to
For these reasons, while it’s often important to have live conversations to hash something out, that doesn’t come in place of writing things down, but complements it.
Okay, but how can we write well? The first thing is to care about it at all. To know that it’s something that matters and that it can be improved, and to start paying attention to it, and to take pride in it. Hopefully I will convince you to do that. Or maybe you already did take pride in it and I didn’t need to convince you of anything.
Regardless, here are some (very) incomplete tips that have helped me improve my writing:
The only kind of writing is re-writing
That’s a quote from Hemingway. Jeff Bezos put it differently:
“Most people,” he said, “think that if they work hard, they should be able to master a handstand in about two weeks. The reality is that it takes about six months of daily practice. If you think you should be able to do it in two weeks, you’re just going to end up quitting.” Unrealistic beliefs on scope – often hidden and undiscussed – kill high standards. To achieve high standards yourself or as part of a team, you need to form and proactively communicate realistic beliefs about how hard something is going to be – something this coach understood well.
When I write code to prototype it, at some point I run the script from top to bottom and voilá, the script runs and the value gets computed. It works! But if you look through the file, there are variables that never get used, sections that pursued some line of thinking that never went anywhere, variable names that are outdated and reference some organizational principle that’s no longer being used, etc. The script is a monument to the bushy tree of dead ends that my thinking has pursued.
Cleaning up this tree (trimming the dead branches) to isolate the successful pathways is critical; the code will be almost unusable to someone else unless you do this. Writing is no different. When you sit down to write something, your thinking will take you in many different directions. Many of these directions will be dead ends, or intersect with something else you’re saying, or take you somewhere you don’t want the reader to go just yet.
So the simple truth is that you simply can’t write something longer than a paragraph in just one pass. At this point, you’re almost certainly beyond the scope of your working memory, and things will come out as gobbledygook unless you edit them.
Partly, this means that good writing is a matter of time and effort, and, as Bezos puts it above: unrealistic beliefs on how long things will take will kill the clarity of your writing.
I like to think of writing as a train of association from what your reader currently thinks to what you want them to think. And if you’re going to take your reader on a mental journey, you should follow many of the same courtesies that you’d extend if the journey were physical:
Meet your reader where they are
Think about what information they’re starting with, what their assumptions are, what their opinion is, etc. Start there. “As you may recall we ended our last meeting with an agreement to do X and Y, because Z; however, I’ve given it some thought and I believe our assumptions about Z were wrong and therefore we should not do X (but we should still do Y)”. If you don’t do this, your reader is going to have to be doing a lot of work to catch up with you — and may guess wrong when they’re filling in the blanks.
Think about where you want to take them
Do you want your audience to be able to run a few docker commands and get something running, or do you want your audience to have a solid understanding of what docker is, why it’s important, and what they can do with it? These are two different places to take your reader, and will have different journeys. Think carefully about what you want your reader to think, do, or be able to do, when they are done reading, and take them there and not somewhere else.
Don’t take them on unnecessary detours
With good editing, you’ll avoid this, since you’ll discard the branches of the tree that are obsolete. Without good editing, you’ll end up taking your audience on lots of random detours that have nothing to do with getting to the end result.
This applies both to the scope of the writing as well as to the writing itself. If you are tasked with making a procedure, make a procedure, not a checklist. A checklist and a procedure are similar, but not exactly the same (procedures have temporal ordering, which checklists do not necessarily have).
In the writing itself, refer to things (people, ideas, past conversations, etc.) in such way that there is no doubt what you are referring to. It is easy to forget how much context you are working with that the reader does not have. Get in the habit of thinking how someone else would interpret your references, and if they might be confused; if so, make your references more specific.
If you say “We should do X” someone can just as easily say “We should do Y”. Now you’re just having a baseless argument, not a debate on the merits. If you say, “We should do X; as we agreed, our main objective for this project is to achieve A, and our alternatives to X are Y and Z. We did some research on Y and Z and they share a drawback P that X does not have.” Now you and your counterparts can surface all the relevant facts and come to a wise decision.
Most of our writing is interactive. JIRA comments, email threads, Slack, etc. Most of what you write is going to be a direct response to something someone else wrote. In these cases, if you want to be a good writer, you first have to be a good reader, and often this means taking the time to really understand what the other person wrote. Ideally, if they have written well, this means that you won’t have to spend too much time on this. But it is important that your response doesn’t completely miss the mark because you’ve misinterpreted what the other person wrote.
Related to this, respond completely. There is nothing more infuriating when you write a message with three points, and your counterpart responds to just one of them without acknowledging the other two. For this reason, I like to respond with the interleaved style versus top-posting, so you can be sure that you’re covering everything.
It is often be the case that you sit down to make a case, and you find that what you wanted to write just doesn’t make sense. No matter what you do, you can’t get to something coherent. This is fine, it just takes some adjusting — you need to convert the purpose of your writing to illustrate why your thinking is confused and to get help. Say “hey, I started writing this to lay out our next steps on Project A, but realized midway through that I wasn’t sure if we were trying to do X or Y. I’m not sure we actually came to a definite agreement on this. Here’s what I think the status is, and where I think we need more deliberation: …”