2018 in review.

By far the most rewarding thing I did this year was marrying a wonderful person. My sister also married, and my sister-in-law is getting married in the spring of next year. In such a sea of marriages, it feels a bit trite to write a year review on my writing hobby, but what’s a hobby good for if not enabling trite behavior?

The other essential bit before I jump into the details is: I’m really grateful to folks who read the stuff I write. I get a lot out of my writing being useful, and always appreciate hearing from folks.

Writing

By volume this was my best writing year since 2008, and I think in terms of quality it was my best year thus far. Looking at Google Analytics, folks spent about 8400 hours reading stuff on my site this year (210k pageviews, 2:23 average time on site), which is just below the number of hours in a year (8,760) – that’s kind of a neat metric to think about. Adding in a 50 minute conference talk presented to an audience of 470, and this year folks spent about a year reading or listening to stuff I created.

I’m sure many folks did far more by that metric this year, but for me that’s pretty cool to think about. Particularly given the audience I’ve been writing for lately – managers and particularly engineering managers – is smaller than some.

The most popular pieces on my blog this year were:

1. Introduction to architecting systems for scale (25k pageviews) - continues to be my best evergreen post, written first in 2011, although I cleaned it up a bit this year.
2. Digg’s v4 launch: an optimism born of necessity (14k) - driven by strong social media and social news. One of my personal favorite stories, and a different kind of writing than I typically do, although I don’t have too many similarly good stories left that I’m able to tell.
3. Notes on “A Philosophy of Software Design.” (12k) - a great book! Honestly, not a very good blog post, it just somehow got picked up on social news.
4. Genetic algorithms: cool name & damn simple (11k) - my second best evergreen post, written in 2009.
5. Migrations: the sole scalable fix to technical debt (9k) - I think the best piece I wrote this year. Has a strong point of view without being needlessly controversial, and I think has actually been useful to some folks. When I posted it, I didn’t think it was very good.
6. Model, document and share (5k) - Shared well on social media, almost entirely thanks to being reshared by Julia Evans ~6 months after I posted it. (Interestingly, this was the piece I wrote immediately after Migrations.)

Of the top ten pages this year: two were written before 2018, seven were written in 2018, and one was a navigation page (specifically “lethain.com/”). To avoid overfitting on initial social media virality, it’s interesting to contrast that against the top 25 pages from the last 30 days, which has a different shape: 7 pages from before 2018, 14 from 2018, and 4 navigation pages.

Altogether, wrote 75 posts this year, which was well over my goal of writing at least once a month, and far more than in 2017 (16 posts), 2016 (13 posts), 2015 (1 post), 2014 (4 posts), 2013 (0 posts), 2012 (11 posts), 2011 (10 posts), 2010 (18 posts). In fact it’s only way back in 2009 (51 posts), 2008 (191 posts) and 2007 (78 posts) that I wrote a similar quantity of stuff.

As part of that writing this year, I got the opportunity to write a book on engineering management, which is tentatively coming out in March. I’m careful about getting my expectations up, so I won’t really believe it’s happening until it’s done, but all signs point to this being a real thing.

Looking forward to next year, my biggest writing focuses are getting this book fully published (!!) and on this blog to write less and of higher value. Roughly the goal is to have more than zero evergreen search-driven blog posts which I wrote more recently than 2012.

Speaking

Coming into 2018, the only public speaking I had done was to fulfill a graduation requirement chapel talk at my high school and a presentation at SF Python in 2013. This year I got the chance to speak at QCon, do three podcasts and one interview that got turned into a post.

In more detail:

• Gave my first external conference talk, speaking at QCon about Paying Technical Debt at Scale - Migrations @ Stripe. After each QCon talk, folks tap their RFID to vote on whether they liked the talk and to get attendance information. This was a bit intimidating for a first time speaker, but was neat to get feedback afterwards and a more precise attendance count (470).
• Did my first podcast ever with Inside Intercom. Not sure how many listens it got overall, the only number I can specifically find is 428 listens on Soundcloud, but that wasn’t the primary distribution channel, so hard to guess about reach.
• Spoke on my second podcast ever on Interestings, which was a discussion of the Migrations: the sole scalable fix for technical debt article I wrote in April of this year.
• Spoke on my third, and thus far final, podcast for Internet History Podcast about Digg v4. This was a reflection on Digg’s v4 launch: an optimism born of necessity, which was my top performing article I wrote in 2018.
• Stripe’s Will Larson on Designing a Performance Management System from Scratch was an article written off an interview I gave based on two articles I wrote (Designations, levels and calibrations and Career levels, designation momentum, etc). It reached number one on Hackernews briefly.

Every one of these was new for me and a great learning experience. Updating my approach a bit for next year: I will gladly accept any offer to appear on a podcast, but probably won’t seek many out. More proactively, I hope to speak at ~three conferences next year, and have already sent out a few request-for-proposal submissions. Most excited for chance to speak in SF, Dublin, Seattle and Singapore, which are all areas where Stripe is actively hiring engineers and engineering managers.

Social media & social news

What I learned this year is that it’s fairly reproducible to write tweet storms that drive high engagement, and that anything of reasonable quality will get likes on LinkedIn. Both of those mechanisms can drive significant engagement. Twitter at its best can drive traffic equivalent to social news, but I think to hit major traffic requires cracking the social news rotation.

Some specifics:

• Twitter - 2,200 up to 3,600. Didn’t get to my goal of 5,000, but I’m ok with that (esp with losing ~500 folks in what I think was a spam follower purge by Twitter).
• LinkedIn - up from maybe 4,000 to 7,400 connections. This was the byproduct of significant time spent cold sourcing, particularly in the first half of the year.
• Social news - two or three posts made it onto Hackernews and did ok there, and the Digg v4 post got some traffic from Reddit.

I spent a reasonable amount of time learning and investing here, particularly in earlier part of the year, and I think my summary is that this aspect of internet publishing is not particularly energizing for me, and it’s something I’m likely to do less of next year. My plan is to not set any goals of this sort, and focus on learning and creating instead – this is, in the end, a hobby.

Altogether, I’m proud of what I was able to do next year, and hope to take some of the momentum into next year while throttling down to something a bit more sustainable for the long-term. Also quite excited to switch from 2018/ to 2019/ directory for the next post.