Before we get into today’s email, here’s a fun thing: I’m partnering up with the consultancy firm Part and Sum to do a survey looking at what’s over- or under-hyped in the world of the tech as we head in 2023. The metaverse, subscription apps, BeReal — I want to hear what you think is cool and what you think is totally over. We’re going to turn it into a white paper which we’ll be sharing back with everyone in a few weeks.

Click here to check out the survey!

I find that looking at data is kind of useless unless you have a couple questions you’re trying to solve. Thanks to internet metrics, you can pretty much optimize whatever you’re doing online into the most popular or streamlined or efficient version of itself if you really want to. But if you don’t give yourself some guardrails I find you either turn into Bored Panda, an SEO farm, or MrBeast. So while I was compiling this year’s stats for Garbage Day, I was looking to specifically answer two questions:

Is my new weekend edition working?
Will I still have a job if Twitter becomes unusable (or I get banned for hurting Elon Musk’s feelings)?

It turns out the answer to both those questions is “yes,” which is a relief. But I also learned a bunch of other interesting stuff this year. Here are the biggest takeaways from the second full year of garbage…

(Lifetime email signups for Garbage Day)

Garbage Day is a newsletter, not a blog

Garbage Day has a total audience of over 39,000, which is 23,000 more readers than where I was a year ago. That's amazing. And my total traffic for the year was a little over five million total views. But, interestingly enough, of the five million total views Garbage Day has generated from 147 issues since December 1, 2021, only a million came from web traffic.

This goes completely against what I had thought last year, that Garbage Day was becoming more of a blog than something read via email. That does not appear to be true. In fact, looking at my traffic from last year, it seems like email inboxes are becoming a more popular way of reading Garbage Day.

There’s one caveat with these numbers, though. Substack, from what I can tell, tracks email opens within a 24-hour period and I publish every other day. Which means, my data is slightly incomplete, which is why I also looked at my total views per post after a week of being published.

Taking vacation is good, actually

My average open rate for the whole year was 50% and my average per month never went below 48%. Which is great for a newsletter of my size and frequency. And I haven’t done any pruning of inactive subscribers, which a lot of larger newsletters do to get a better open rate.

My biggest dip in readership this year came when I took a two-week vacation in August, with total views and email opens dropping 33% and 27% respectively. But my “LCD Soundsystem reunion theory of online publishing” that I wrote about in last year’s yearly traffic report is still accurate. When you come back from time off, you bounce back hard. The views and opens in the month following my vacation were up 57%. My total traffic for September was 100,000 views higher than where I left it in July.

Taking vacation is just good business sense. Which is why, starting later this month, if Garbage Day falls on a holiday, I’m taking it off (this won’t apply to content for paying subscribers (for now)).

Ask readers what they’ll pay for

A free version of the new paid weekend edition was the largest driver of paid subscriptions this year and the weekend editions are being read more regularly and converting more customers than anything I was trying before.

The idea for the weekend edition came from the big reader survey I conducted in August. The number one thing people were asking for was just an easy way to cut through the chaos of using the internet. So I ran with that and it’s worked like a charm.

I also tried an experiment with partial paywalls this year and I can report that it was a flop. Substack launched the ability to put a paywall in the middle of a newsletter, so I tried it in 18 issues this year. At the bottom of a free issue, I would put extra content for paying readers. Not only did this negatively effect the engagement and performance of those issues — on average they performed 18% worse than both normal and fully paid issues — they also didn’t really convert anyone. (I suspect part of the reason for the poor performance is that if you do that, the whole post gets a little lock icon at the top and makes it look like you can’t read it unless you pay.)

In terms of the business of Garbage Day overall, I'm still not a Substack millionaire. And I wouldn’t say I’m making enough that I can fully unclench my jaw, especially as we head into a possible recession, but I can pay the bills. I also set a reasonable financial goal that I figured would be a good benchmark for determining if I could keep going another year and I hit it! So I’m pretty proud of myself.

Twitter is a (useful) illusion

My social strategy for Garbage Day is what I call “the one good paragraph”. I try and get something punchy and retweetable condensed into a single paragraph somewhere in the newsletter so that I can screenshot it and share it on Twitter. And a lot of times my feelings about how good or bad Garbage Day is performing is based on the engagement of that paragraph. Turns out this is a bad idea and a massive waste of my time (sorta).

Twitter was my fourth largest traffic source this year. That sounds like a lot, but that only amounted to around 100,000 total views. That’s 2% of my total traffic! Even worse, Twitter users don’t really sign up and definitely don’t pay. Less than 1,000 users who came from Twitter signed up for free emails and less than 80 paid me. In comparison, 2,000 users from Google signed up for free this year and over 100 converted into paying customers.

Now, do I think it’s worth completely throwing out my “one good paragraph” strategy? No, I think it’s important to think about how someone will share your work and you have to bake those shareable moments into the work itself as you’re writing. But I think it’s time to start really decoupling my sense of success on here from retweets.

It’s worth explaining memes in a timely fashion

Two out of the top 10 Garbage Day posts by total traffic this year cracked 70,000 total views. And I’ve been trying to figure out why:

My best guess for “We have too many Main Characters now” is that it was just something people were really excitedly forwarding to each other. According to Substack’s metrics, all the traffic was from email and I can’t find anything public pointing to any one specific thing in the email that people really liked. But, hey, I’m not complaining.

As for “The truth behind Finland's ‘catgirl’ prime minster,” I know exactly what happened. According to Substack, 35% of the traffic came from Google. I assume because people were googling “Finland,” “prime minister,” and “catgirl”. Because you’re all perverts.

People open emails that feel like emails

What’s interesting about the top 10 most-opened Garbage Day emails is that seven of these posts didn’t crack the top 10 in terms of total views:

We have too many Main Characters now
It's ok to feel sad about Twitter
One giant slow-motion fail whale
All news is good news in an attention economy
The TikTok slot machine
Men too weird to just read Barstool Sports
Someday soon, you will tweet for the last time
Social media is just a thing that happens
Elon Musk is more Funko Pop than man
A power grid for the whole world

So why did these do well in inboxes? First, and most superficially, five of the top 10 most-opened Garbage Day emails had around 32-36 characters in their subject line, which means my discovery last year that 39 characters is the optimal length for a subject line is still correct. But I also don’t think short subject lines or sticking emojis in them or whatever is enough to really matter long-term.

The deeper takeaway is that there is a subtle, but noticeable difference in tone with the subject lines. The posts doing the best as emails tend to be written more, well, like emails — thoughtful, contemplative, eclectic, conversational. I’m a firm believer in content being made for the distribution platform it’s going out on, so it makes sense that people like to open emails that feel like emails.

People like emails about different topics

I’ve said this before, but I don’t look at my unsubscribes, or “churn,” as it’s called. My emotional constitution is simply too delicate to handle it. (Unless you email me angrily and usually I’ll email you back.) But I also think trying to find out why numbers are going up is the same, if not more helpful, as trying to find out why numbers are going down. And so what stands out when looking at the top 10 Garbage Day posts that drove the most free email signups is that they’re very diverse thematically:

Many of them have slightly out-of-left field opening essays. And many are focused on platforms other than Twitter. I assume it comes down to variety, a sense that if you sign-up you’re not just going to get emails about the same thing over and over again.

(Month-to-month percent change for views, email opens, and average open rate)

People will pay for snark or a service

The week of November 7th was the largest conversion of free readers to paid and the second largest conversion came during the week of May 23rd. And four of the posts in the list below of the top 10 posts responsible for the most paid subscriptions were published during those weeks:

I don’t think there’s anything especially interesting about those two weeks, but I do wonder if people are more likely to pay at different times of the year. After looking through all these metrics for a second year, I’ve noticed that there’s a real newsletter malaise that happens between February and April, which recovers in the late spring, flatlines across the summer, dips around early fall again, and starts to recover as we head into the winter. Though, I may have also just accidentally mapped out the seasonal quality and enthusiasm of my own writing. Who knows.

But I would also say that people tend to pay you for something servicey, like the weekend edition or the redesign of my Discord, or because I wrote just an absolute barn-burner of a hot take. Which is not something I can or want to do all the time. So it’s good I have options.

Lean into what works the way you want it to

As you can see, there’s a million ways to optimize a newsletter. I could race to build a gigantic free list and focus on advertising. I could add more paywalled issues and really cultivate my paying audience. I could really aggressively target Google traffic. I could become a blog and stop worrying about email traffic or vice versa.

I need a few weeks to really digest what all these numbers are telling me, but I think it ultimately comes down to figuring out what it is you’re good at and like doing on the internet at right now, what it is you want to be doing a year from now, and picking the avenue that will get you there in the time you want. A little simplistic, perhaps, but it’s the easiest way I’ve come to think about *~content~*

I’ve got some fun ideas for next year and I think I’m mostly in a place to try them. And I’m really grateful I get to keep doing this. Garbage Day is, without question, the best job I’ve ever had. But I’m definitely taking more vacation next year.

My last live event of the year is next week! It’s at Caveat in New York on December 10th. It’s going to be a blast. Our speakers include Jordan Uhl, Abby Govindan, and Manny Fidel. You can pick up tickets here.

P.S. here’s a good idea for next year’s World Cup.

***Any typos in this email are on purpose actually***

Garbage Day

The Big 2022 Garbage Report