An Iterative Approach to Writing

time spent writing


The waterboarding of the ego. The crushing of the soul. The realization that I can't spell.

I'm talking about writing.

I like it but it doesn't come easy. It's hard. Kind of like running a marathon is hard. Except with running, I've learned to enjoy the struggle. With writing, this is still a grind.

I'm a sporadic writer - like sunshine in Seattle, writing comes to me only a few times each year. I tend to write when I need to rant or feel the need to explain something.

When I do write, I always feel that I spent too much time on it. I feel as if I'm doing something wrong, that it shouldn't take as long as it does.

I figured I couldn't be the only person feeling like this so I've decided to share my writing process. Maybe you're like me and this can help put your mind at ease. Maybe you're not in which case you can give me pointers on what you're doing differently. In any case, this is a case study of my writing process.

CDK All The Things: A Whirlwind Tour

This was an article I published about the AWS CDK when it was still in preview. I just started using the AWS CDK at the time. I was a recovering YML addict and a cloud formation junkie. The CDK felt like the answer to a prayer I didn't know I had.

My strong feelings for the CDK led me to write this article.

Multiple Versions

I write in versions. I start with an outline. Then I create a V1 (version 1). Then a V2. And so forth. The CDK piece had three versions though sometimes I can end up with over six versions by the time I'm done.

I start with the outline. This goes over what I want to write about using a few bullet points.

  • how I started using it
  • goal of article
  • overview of cdk
  • usage experience
  • verdict
    • love
    • good
    • edges
  • takeawy
  • next

V1 comes after the outline. This is my version of the shitty first draft (Private). In V1, I add more details to each of the bullets and the overall structure of the article starts to take form.

V2 is when real sentences come in. I tend to start from scratch here and write sentences and transitions from the bullet points in V1. I will add titles of articles that I want to reference and leave a TODO for the actual link. In this version, my main focus is getting coherent prose down without being interrupted. If I'm stuck on the right wording for a particular sentence, I will leave a TODO there as well.

V3 is the first editorial pass. At this point, I fill out all the TODOs and go over each sentence, reading it out loud to myself to hear how it flows.

If V3 still needs more work, I will do a V4 and potentially additional versions until I feel like the piece is ready. At that point, I will copy and paste everything into grammarly and do a final pass for grammar and spelling before publishing.

Multiple Phases

I like to write in dedicated chunks of time. I aim for at least 30min and if I get stuck after that, I will usually stop and work on something else and pick up writing the next day. The below is my writing log for the CDK piece.

  • 1

    • 2019-06-09T07:12: start
    • 2019-06-09T09:21: stop, total: 129min
  • 2

    • 2019-06-10T16:52: start
    • 2019-06-10T17:55: stop, total: 63min
  • 3

    • 2019-06-11T10:20: start
    • 2019-06-11T12:17: stop, totoal: 117min
  • 4

    • 2019-06-11T19:07: start
    • 2019-06-11T20:12: stop, total: 65min
  • 5

    • 2019-06-12T14:50: start
    • 2019-06-12T15:36: stop, total: 45min
  • 6

    • 2019-06-13T08:55: start
    • 2019-06-13T09:42: stop, total: 50min
  • total: 7.8h

In Closing

All in all, the piece took me 8h over 4 days. You can find all my intermediary work on the piece here. The final output is 3.5k words or 500 words per hour. To put that in context, I was doing independent AWS consulting at the time and 8 hours equalled a few thousand dollars in missed billable hours.

This is another reason why writing was hard - any individual session would rarely be worth the missed earnings. That doesn't mean you shouldn't do it - over the long term, I find that it's one of the most effective levers to increase your luck surface area (Private). Once you publish something, it will exist forever (like plastic), earning SEO points and potential new relationships. Not taking time to write because you have other things going on is the true opportunity cost.

cavemen too busy

The piece mentioned in this case study did have a happy ending - it was picked up by someone from AWS a few weeks later and AWS became one of my consulting clients. If you are interested in the process that went behind this current piece you're reading, you can find the notes for it here. Happy writing!


  1. pkm (Private)