Hey everyone. My habit of writing regularly has completely disintegrated, and I'm in the process of developing a new workflow, but in the mean time I thought I'd say "hey everyone" and give a quick update.
Probably the most significant thing that's happened with The Sample is that I spent a week and a half overhauling the ad system, and that took effect Thursday morning. Now, instead of throwing a random link at the top of the emails, I'm turning 10% of emails into sponsored forwards. Newsletter publishers can bid on the forwards, and they pay whenever someone hits the 1-click subscribe link.
Whenever someone subscribes from a paid forward, I get an email alert like this:
It's averaged about $21 per day so far. I expect it'll be higher during the week, especially Monday.
So that's great news--even $20/day would be $7,300/year. That's not huge, but it's enough that being able to support myself with the business no longer feels so distant and amorphous. I can see it on the horizon now.
Next up, I'm doing some more improvements to the algorithm to make the recommendations feel less random. Right now people often get the feeling that we aren't really taking their selected preferences into account. We do in fact, but it's fuzzy and probabilistic. I think people assume that we have hard categories for each newsletter, whereas we're actually just using a language model to calculate distances between keywords that people select and keywords for the newsletters.
Several weeks ago I added a tags field to the rating page so that we can crowd-source the newsletter tags instead of trying to extract them automatically (which was a big contributor to the fuzziness). My wife and I have also gone through all the newsletters and seeded them with tags ourselves, though to be honest I'm not confident in my ability to come up with good tags--I'm happy that it's mainly crowd-sourced. Now that we've collected a bunch of tags that way, I'm going to make changes to the way we use tags.
After that I'll likely focus on growth. Maybe write some blog posts and include them in the weekly announcement emails that go to publishers, so they have more ways to share The Sample if they want.
If you want to get more commentary like this, you should follow @the_sample_umm. To simplify things I've decided to just focus on Tweeting things from that account. (And to the ~100 or so of you who subscribed to this because I linked to it at the end of The Sample's signup flow, sorry for the switch-a-roo!). I am going to continue writing about "the future of discovery" in general in this newsletter, eventually, after I get a good workflow set up. Part of that involves writing some more software. See this thread:
In other news: I (we!) have slowly started to work on an app I'm codenaming "Infovore." it's the nth iteration of https://jacobobryant.com/p/we-need-user-centric-data/. it'll include a bunch of stuff (rss reader, bookmarks, email client, blog, newsletter, tweet scheduler...) all mashed together, ...
... intended for power users. It'll be self-hosted + plugin based. Think "vim/emacs for reading and publishing." Imagine if the way we consume and share information was as individually streamlined as the way we edit text.
When that's figured out, I'll also start sending regularly from my Software Invention newsletter, which I created a landing page for but still haven't sent out a single time.