I just sent the last Squirrel Notes of 2018 — Issue #20 — which concludes the first full year.
If you’re not familiar with Squirrel Notes, it’s an email newsletter about content management that I publish about twice a month. Each issue has five “notes,” which are just random things of interest that I’ve picked up from the world of CMS.
Here’s a little self-examination about why I started this, and where it’s going.
An email newsletter seems like such a step backwards. But there were a few things that struck me last year —
I subscribed to a newsletter called Recommendo, which is nothing more than six personal recommendations and tips from a group of editors. It’s quirky, and idiosyncratic, and it directly represents nothing more than the genuine opinions of actual humans. I love getting the newsletter — meaning, I literally smile when I see it in my inbox. I wanted to do something that would (hopefully, fingers crossed) give someone else that same feeling.
Just as much as Recommendo, I love the car videos of Doug DeMuro. He has a YouTube channel with millions of subscribers, and he reviews cars — new cars, old cars, all cars. His videos are about 25 minutes, and they’re not…corporate. When Doug reviews a car, he concentrates on what he calls the “quirks and features” of the car (he says that so much, in fact, there’s actually a drinking game).
One day, while watching a video, I realized this is what I identified with — the weirdness, the intricacy, and the specific oddness of what sets one car apart from another. Doug makes car content for weird car people, not normal people. He talks about random, odd things that only occur to people who really love cars, and I started wondering if I could capture this same vibe for people who love CMS.
(I think the word for this is “wonk.” A wonk is someone who geeks out on the esoteric, intricacies of stuff, sometimes with little practical application. Sara Wachter-Boettcher once called me a “CMS wonk” at a conference in Helsinki. I’ll own that.)
Also this year, I was traveling a lot and talking to a lot of people about CMS. I love it when two people who enjoy CMS start rambling about it, and topics just start falling from the sky. This person knows that person, and this system is going in that direction, and this trend is about to become the Next Big Thing. There’s so much information that just drifts by, and I wanted some way to call it out for people who are as weirdly interested in it as me.
So, I had started taking notes on random tidbits I picked up on the road (read: “notes”), and then this collided with my aforementioned love for Recommendo, and…
I don’t know at what exact point I took the actual plunge, but this tweet is from January 8th.
I have this urge to create a monthly email newsletter about CMS. I have no idea why I want to do this so much.
I put together a trial issue and then ended up writing another 5-6 in advance. I was actually sitting in a CMS selection process next to Karen McGrane in NYC, and during breaks, I would send her test emails of various issues. She encouraged me to forge ahead with it.
The only name I ever considered was Squirrel Notes. My book has a squirrel, and I was emailing my notes, so….well, there you go. The entire idea is literally notes.
I made a couple decisions up-front.
- I wouldn’t ask for any information other than email. I have no plans to market to the list, so I didn’t need any other demographics — hell, I’m just happy you’re here. An email address often reveals identity and employer, but subscribers are otherwise anonymous.
- I would try to grow the list slowly and organically. I don’t want anyone on the list that doesn’t really want to be on it.
As of this writing, the list is just over 300 subscribers. And I’m proud that only about 5% of people who have ever subscribed have unsubscribed. So, only about one person in 20 thinks, “Nah, I don’t want to read this after all.”
(That math, BTW, is: total unsubscribers / current subscribers + total unsubscribers.)
I haven’t done any marketing outside of some Twitter posts and a LinkedIn post or two. All the growth has been word of mouth. I’m hesitant to do a big marketing push because I don’t want someone to subscribe who isn’t genuinely interested. Also, the newsletter doesn’t really have much of a commercial point to it. I’m sure there’s some incidentally marketing benefit to me and Blend, but that’s nothing you can calculate the ROI of a marketing campaign around.
(That said, if you want to mention Squirrel Notes to other people you think would be interested, please do.)
To let people know what they’d be getting, I have a public archive of most all issues. I hold the latest two back (otherwise, why would you subscribe?), but at any given time, you can see any other issues. I’d rather a potential subscriber look through some issues and make an educated decision to subscribe, rather than subscribe blindly, get an issue or two, and decide it’s not what they wanted.
(This just happened the other day. A guy subscribed, got one issue, then unsubscribed, which sadly inflates the stat I mentioned above. This annoys me.)
My feeling is this: there are only X number of people in the world that would be interested in this over the long term. What is that number? 500? 1,000? It’s certainly not much — of the emails I recognize, I can tell you that the list is basically consultants and vendors; so “industry insiders,” to some extent (“wonks”!). That’s a fairly small group, and I don’t think a contrived, artificial list of 20,000 vaguely interested people is going to do me much good.
I monitor bounces carefully. Curiously, it tells me a bit about comings and goings in the industry. I see a bounce which strikes me as odd because I recognize the email, and then I’ll see a notification that the person has switched jobs.
Also, it’s very gratifying to see that person resubscribe under their new email or their personal email account, which tells me that they want to make sure they keep receiving the newsletter (in that case, I delete their old email from the list, so it’s not technically an unsubscribe).
I sweat the open rate. MailChimp tracks how many people open the email (as best it can, given privacy settings). The list currently has an open rate of over 50%, which is very good, and there’s likely another 30% of people who prevent their email clients from registering the open (and this is likely even more common among my tech-savvy subscriber base). I’m reasonably confident that 80% or 90% of subscribers read each issue.
If the open rate were to drop precipitously, then I’d have to give serious thought about whether to continue, for this simple reason: if it’s not a thing someone wants to see in their inbox, what’s the point?
Every issue after the first two has five notes. Those first couple had six, because I had no plan and I was just throwing in all sorts of stuff. I quickly settled on five as the standard.
I’m pretty ruthless about word counts. Very few issues will have over 600 words. A college-educated adult can read about 350 words per minute, and skim much faster. My goal is to make Notes a two-minute read at most. If an issue is “heavy,” I will trim pretty ruthlessly — delete entire paragraphs, remove wordy clauses, or just nip and tuck everywhere.
I saturate each Note with links. My goal is for you to click something— I want you to see something interesting and investigate it. My mission is basically to write teasers to get you interested in other content. I want to “incite exploration,” to put a dramatic phrase on it.
I stay pretty close to CMS, but I edge into content technology in general. Really, I’m interested in anything related to CMS that would broaden the perspective of somebody who works with this software for a living.
I used to put the content in categories — things like “theory,” “events,” etc. When I was sending her test issues, Karen told me to dump the categories as they didn’t matter. But I was smitten with them, so they stayed…only to get dumped a few issues in, when they got tedious. She was right — you can categorize them in your own head; you don’t need me to tell you where they fit in.
Also, I used to show teasers at the bottom — “In the next issue of Notes…” I did this because I often had the next issue partially written as the current issue was being sent…but not always. It got to be a bit of a pain, and I was often thrashing the next issue together just to have the previews. So, I’ve dropped this for now. It might come back.
I try to get one issue out every two weeks, but not always, and I’m okay with that. For Notes to have longevity, it needs to not be stressful. So if I’m busy on a given week, I just wait until the next week. I decided early on that I wasn’t going to stress about it. I know you’re not staring at your inbox waiting for Notes to arrive, and a self-imposed schedule was just going to make me resent it.
I’ve never put any effort into the timing of delivery days or times. I send it whenever it gets done. If it’s outside working hours in North America, I might schedule it for the next morning, but I’m all over the map here.
I wrote a book on CMS, so you know the technical process behind each issue of Notes is stone-cold dialed in to the best possible industry practices. Woot! Woot!
It’s not even content-managed. I write each issue as a delimited Markdown file. The files are sitting in Dropbox (where they are versioned), in folders by issue number (“001”, “002”, etc.). Each folder has a “notes.md” file, and a “header.md” and “footer.md” file if I have anything special for the header and footer of that issue.
I have a C# script that I run directly out of LinqPad to create the email. All the files are parsed and passed to a DotLiquid template. That template generates HTML which is pushed into MailChimp via their API (I have never used the actual MailChimp editor for Notes).
When MailChimp is updated for the first time (so, when the campaign is created), the script writes an “issue.meta” file in the same folder which contains the campaign ID (so I can update it in future executions). After the campaign is actually sent, I run the script once again to get the number of emails sent and the date they were sent on, which get written to the same file. This is the the metadata that powers the archive.
The script also generates a flat HTML file for the archive, which is written to a private “server” (in quotes because it’s actually the computer sitting under my desk in my office). The website (which is out on Azure) calls to it for an issue it doesn’t have, then holds the issue in cache. When the script runs, the last thing it does is call an endpoint on the server to clear that cache, so any new requests will get re-fetched.
Therefore, my “process” is simply this: I write, execute, review, write, execute, review, and so on. It’s awfully low-tech, considering what I do for a living.
But I have no plans to change the process, just the output. I’ll always compose the Notes in a Markdown file, but I’d like to eventually push them into a hosted repository of some sort, for reasons I’ll explain below.
The technical part of Notes has been really fun. It’s low-tech, but it’s not hacked up. The code is solid, and I actually haven’t touched the script or the template in months. It works beautifully.
It’s also a neat example of doing something “exploratorily.” I had no idea what I was going to do originally, so I didn’t want to put a lot of effort into it. I figured I’d screw around and see what happened. I had the following luxuries:
- I’m the only editor
- I’m technical, and comfortable in Markdown
- I could template them myself
- Delivery is decoupled — MailChimp and static HTML are my delivery channels
When I do finally decide to model them for a repository, I’ll be in a great position because I’ve been tinkering with the process and format for a year, so it will be nicely established by then.
Note that I surely wouldn’t do any of this for a client, but for a hobby, I’m having a blast with it.
As I mentioned, one goal is 2019 is to move the individual Notes into a content repository of some kind. Through 20 issues, there’s just over 100 of them now (not exactly 100, because of those pesky six-note issues in the beginning).
I’d like to enhance the archive. Right now, you can scroll through issues, and view a full issue. This list is getting longer, and it’s eventually going to have to paginate, which gets weirder with flat files.
Additionally, I’d like to:
- Give each note it’s own permanent URL
- Make the archive searchable
- Tag posts, to group them by subject/concept
- Be able to link posts back to each other
That last one is important. This last issue, for example, I made an addendum to a Note from the prior email. I’d like to refer to that prior Note, but I need a URL to do it, so they’re eventually going to need permalinks.
Additionally, the tagging would enable me to track “conversations” that develop over time. I remember this need from my blogging heyday — I’d do multiple posts on the same theme, and I wanted a “topic page” that recorded that particular conversation. (Here’s a semi-cringey post about this from 13 years ago: Coming Soon: Guided Content Maps. I never did do it, though.)
Finally, one other thing I need to do is modify the subject lines. Right now the subject is just “Squirrel Notes #20” or something. I’d like it to still have that standard, but perhaps have a slug from the lead Note to give you an idea what it’s about. I hope this would increase open rates, but there’s a chance they might get worse (what if the slug isn’t something you’re interested in?).
So, that’s it. I’m one year and 20 issues in, and I hope to continue it for as long as possible. I hope you enjoy seeing it in your inbox every couple of weeks.
- Read my first book: Web Content Management: Systems, Features, and Best Practices
- Read my second book: Real World Content Modeling: A Field Guide to CMS Features and Architecture
- Subscribe to updates from my next book: The Web Project Guide
- Subscribe to my twice-monthly newsletter about CMS: Squirrel Notes
- Follow me on Twitter, where I announce new posts: @gadgetopia
- Send me an email — I'd love to talk: firstname.lastname@example.org