2024 Client/Server year in review
I've only operated Client/Server for a month, but it's time for the inaugural Year In Review recap!
First of all, thank you to everyone who has subscribed! When I had my daughter almost 20 months ago, she became my entire life and I did not have much additional time for creative pursuits outside of work. But she is sleeping well, life feels more stable, and I finally have time for a hobby. I’m so grateful to everyone that has read my posts, and I am also grateful to everyone who has reached out to chat about my newsletters.
I’d like to end every year with a recap. Even though I’ve only been writing the newsletter for a month, I’m going to recap it!
Top content of 2024
This is going to be short since I only launched at the beginning of December, but my 3 best performing pieces are..
Spotify Wrapped fell short this year. This post draws a straight line from this disappointment to the layoffs they had last year.
I’d like to write more pieces like this! It was timely; I actually axed a different story to work on this and had a lot of fun doing it. I’ve also been in the software industry for a long time. This gave me some extra perspective on the “how the ___ did they ruin their unique experience?” aspect of this story.
Redis is using its trademark rights force changes in client libraries, with the alternative of strongarming third parties into handing over control of the libraries to Redis. I also explain how developers end up in these situations by exploring a library that I had published, a Gravatar library.
I enjoyed digging up my old Gravatar library. Part of me wants to revisit the library and release a 0.0.2 with the now-larger image size that Gravatar supports.
This post explores just
and other alternatives to Makefiles. It also explores why Make has lasted 50 years and what it might take to unseat it.
I was really happy that this post got some traction. It validates my feeling that I can mix current events and nerdier topics. Make has also been a continual thorn in my side for the past 20 years and this was therapeutic.
What went well in 2024?
Cadence
Two posts per week is sustainable so far! It’s definitely the best cadence for my schedule right now. I have to write and edit 4 or 5 days a week, but I don’t feel rushed when I skip Thursday and Friday night to play video games.
Content
I’m still searching for my exact formula. But I’m proud of most of the newsletters that I wrote, and I’m glad that the month validated my initial hypothesis that I can mix current events and in-the-weeds pieces.
I also saw a pretty direct correlation between effort and quality. Effort also tended to correlate with more subscriber traffic. Is this cause or effect? It’s hard to say; it’s believable that I did a stronger job polishing and promoting pieces when I was happy with what I wrote.
Self promotion
I’ve always felt awkward about self promotion. I feel uncomfortable posting most of my newsletters to LinkedIn. However, it’s also helped people in my professional network discover it. I’ve also had some delightful conversations as a result of people coming across it on LinkedIn. So I’m going to continue this for the time being. It is Linkedin, after all.
Planning
Have you ever wanted to launch a new project, only to be delayed a day or two by buying the domain, setting up redirects, debugging why SSL isn’t working, etc? Well, I’ve done this enough times that I actually did all of the setup 3 days before I wanted to launch. Look at me, finally learning a lesson!
Areas to improve next year
“Stretch singles into doubles”
Nate Silver had a great post called “Always. Be. Blogging.” containing his recommendations for running a successful Substack. The point where I suffer the most is probably “Stretch singles into doubles,” i.e. actively building and marketing the content to promote itself. Since I launched at the beginning of the month, I’ve only put cover images on three of my newsletters. But 2 of the 3 were well above average for views. Nowhere close to a statistically significant sample, but it is interesting!
I have a few takeaways for 2025:
Look at what other Substackers do. I have not done this.
Research how to write headlines and ledes. Surely this is combination art and science. For example, I’ve already noticed that the subject of the piece should be in the headline.
Experiment with different social media promotion formats and see what works.
Make a “launch post” checklist incorporating my lessons.
Slow news weeks
The news started slowing down in the third week of December. I wasn’t prepared to have a slow news week so early in the publication’s life. So I scrambled to generate one or two post ideas.
In retrospect, I should have been prepared for this! I’m always bored out of my mind the last 3 weeks of December because there’s not enough Content on the Internet. And I can prepare better in the future. I have a few options moving forward:
Not-so-current events. Interesting topics that were bumped by more-important news. This did eventually occur to me, and my Markdown piece was a good example of this.
War stories. I like incorporating war stories into my posts when they are relevant. But I have a few good stories that will probably never get slotted into a post. Like the time I destroyed $10,000 of computer equipment in a shower of sparks with a car battery.
Followups to previous posts. “Where are they now?” or “The rest of the story”-style posts.
Rants. I’d like to avoid these. It’s not the tone that I’d like to set. But on the other hand, would it also be fun to spend a few hours ranting about Bazel? Would it be fun to write a post titled “JavaScript Isn’t Real”? Probably!
Things to remember moving forward
Most — but not all! — of my subscribers are people that I know. So I have “impossible” metrics right now like a >60% open rate, and I can likely expect that to drop far south of 40% when I continue publishing.
I should expect a low subscriber counts for at least 6-12 months. It takes regular brand recognition to build subscribers, and Client/Server is new and does not have an online following. I’ve also heard in SEO circles that it can take Google over 6 months to start recommending a new domain.
There are zero words of LLM-generated text in my posts, and this will always be true. However, I have not spent enough time researching and strategizing with different LLMs.
Goal for next year
At the close of the year, I ended with 21 free subscribers and had 576 30-day page views. My stretch goal for next year: end the year with 1024 free subscribers. I’ll be lucky to get 10% of that number, but setting an aggressive goal will force me to think harder about how to achieve this.