2.0; DjangoCon; Investments vs. bets
Lord forgive me but it’s time to go back to the old me.
— Bugs Bunny
I can think of no better way to revive a years-dead newsletter than with a little bit of navel-gazing, so strap in because your appetite for five minutes of me dithering and rhapsodizing about content strategy and identity will be a useful bellwether for the things to come.
One of the interesting things about Buttondown as a product is that we've really elevated a use case for which I'm still trying to find the right phrase, but for now I call it "headless mode." What this means is you can basically set up Buttondown, connect an RSS feed to your newsletter, fire up a good old-fashioned form tag, and then you're sort of done. Buttondown will work away in the background, sending a digest email every week to your growing subscriber base. This is really nice and, in many ways, the acme of what I originally wanted out of the tool when I started it so many years ago.
The problem, of course, is that you don't log into the tool and you don't use the tool (or at least "the tool" in a very conventional sense of the word), and suddenly your touch sense of the quirks and rough edges and places where users of all ilk stub their toes dims and fades away. This is pernicious, especially because headless mode—aka the way I use Buttondown through jmduke.com, a very lovely personal web page, which you should still subscribe to if you haven't already—is how I primarily interact with the product. For the vast majority of users, the two biggest product surfaces are the editor and the archives. These are also the two surfaces that I do not use on a day-to-day basis.
Thus, the revival of weeknotes. I use the word "revival" very specifically. One or two of you might recall that literally five years ago, I kept week notes where I would write about everything from architectural changes I was mulling to pricing changes and all of the really boring stuff in between. Those writings are still public (check the archives!), and they will be so forever. But it's hard to return to that exact style and cadence: the reality is the earnestness and brashness with which I was able to treat a project that I cared deeply about, but a side project nonetheless, cannot be extended into 2025. It feels uncomfortable to talk about many things. I know that speaking or writing poorly can, in some minute way, jeopardize my ability to pay myself and my team health insurance. So a lot of the more interesting writing has been in private to Buttondown folks, to friends. This is another great use case for Buttondown private newsletters.
But I think I've swung a little too far in that direction. It's true that there's a lot of stuff that I can't share. It would be disingenuous of me to pretend otherwise, but there's still a lot I can share.
So, weeknotes 2.0. Here we go.
Last week was a combination of firsts. Buttondown sponsored DjangoCon and went in person to man the booth—the first time we've ever done something quite that grand. We've sponsored a couple conferences before, most recently BangBangCon, but that was all remote and this was a thing that involved tabling and badges.
Second, we had our first offsite. Four of us, myself included, all collided in what was frankly a very lovely WeWork, converted from an old bank building, for eight hours and hammered out some of the big existential questions we had looming. These two existential questions are probably unsurprising ones, depending on how much you've listened to me rant about work over the past year.
They are, in order:
The quality and reputation of our little tool has increased pretty dramatically. And as a result, we've started to gain a lot of users who we wouldn't consider our ICP. Who are we building for, specifically?
With that in mind, what are the highest leverage projects that we can work on over the next six or so months to serve those users?
Now, the thing about questions like these is that the null hypothesis is not a failure case. Sometimes it can not just be useful, but vital to exorcise the demons of paths not traveled. And lest I bury the lede, that's pretty much where we ended up, and quickly at that. Even if our median new user is slightly outside the confines of who we think of as our “ideal user” (put plainly: lots of people who choose us as “Substack, but different” or “Mailchimp, but different), we’re not expanding our company’s strategic Overton window. We’re building with our very specific, idiosyncratic audience in mind: if the artifacts of that work continue to resonate more broadly, lovely, but that’s a second-order effect. (Think Friedkin following up The Exorcist with Sorceror.)
In particular, where we quickly turned our attention in terms of project work were some of the more death-by-a-thousand-paper-cuts experiences that exist in our app today. Here's an obvious one: double opt-in. Double opt-in has existed for literally the entire lifespan of Buttondown. It's a very well-paved piece of user experience territory, and there's not much that can be said about it besides that the main goal is to not fuck it up. And yet, bit by bit, piece by piece, as the extended universe of Buttondown has built, grown, and complected around double opt-in, that core flow has suffered. The way we handle upselling opportunities is incoherent. The way we do white labeling and branding is inconsistent with other transactional emails. And in general, going through the opt-in flow as anything other than a very, very standard case newsletter has at least 2 or 3 bits of room for improvement.
When I was telling a friend about this, he asked in a joking but not joking way, "How does this impact the bottom line?" And the joking but not joking answer I gave him was a simple one: It does. And that's more important than knowing how much.
I'm north of six years into growing Buttondown, and this is the sixth year that we will have either doubled or come close to it year over year. In these six years, I can't point to a single instance of trying to really, really lean on numbers for deciding what work is worth doing. That won't last forever, nor is it a value judgment against those who do. But I tend to think about the work we plan and partake in as falling into one of two buckets: investments, which are lower risk and higher yield over a longer time horizon, and bets, which are comparatively much higher risk but may have a higher return, especially in a short time span.
And this framing — investments over bets — is one that I leave Chicago clutching, even if it leads us in the same direction that we were heading anyway. There is still so much obvious work to do.