You’re Already Optimizing for Attention

The modern "attention economy" can be brutal for everyone, but I think musicians are especially exposed. The very nature of performing music for other people is essentially making a time/attention demand of the audience. Always has been. It's a challenge you have to adapt to. But the musical world of social media can be very saturated at times, and you can't really afford to avoid it unless you want to be invisible.

I also think there is a weird paradox: It seems like any social platform can be fairly saturated when you're trying to get your work out there, but from a media consumer perspective, a lot of what we end up seeing isn't the actual music or even related to promoting any musical recordings/performance. It seems like another version of the same economic forces that led MTV to pivot from music videos to documentaries about trying to hook up in New Jersey. Classic. And I don't mind when that kind of content is packaged in the medium of a Bon Jovi album, either.

Recently, I posted an article about using Reverb's API to automate parts of the process of selling guitars online. While that article focused on just one API, it was clear that I was starting to see the need for social media automation in that post too. If you have guitars sitting around on reverb for a long time, doesn't it make sense to have some sort of automation that would automatically post to various social media sites based on the actual listings in your Reverb account? And once we are in the land of talking about social media APIs, there is enormous, 2-sided value to musicians. You can save time by both automating your social media and analyzing it to get data-driven feedback into what works (and doesn't).

I'm not a social media expert by any stretch of the imagination. In fact, I have been outright terrified of it for much of my adult life. But in my day job as a "web application architect" I am paid to engineer growth in reach and visibility for my company's ecosystem and solve problems in ways that scale to entire industries. In that setting, a lot of my engineering work has gone into SEO tech/automation. For most musicians though, social media is the more important visibility opportunity. And I think it can be systematized through automation and analysis. It might sound like overkill at first, but I'd argue that it's part of your job/career if you're a professional musician (I am not one), and it can help buy back time too for practicing, writing, recording, and more.

One of the big themes from my very first post here (which actually inspired me to build this whole site) is that companies like Fender and Reverb flood the zone with ads and intentionally need to do so for their ads to have an impact on buying. And major legacy bands pay people/companies to manage their social media accounts. We don't want to lose authenticity, but I argue that you have to embrace automation and a data-based process to keep up in a world that can eat all your time and attention without even giving you a chance to pick up your guitar if you aren't careful.

Not all platforms are the same

Even when it seems like 2 platforms are basically the same media format with different user bases, the experience as both a user and poster can be totally different. Tiktok !== Youtube Shorts and Twitter !== BlueSky.

BlueSky from my experience (for both this site's social media and my own personal social account) is pretty strong in terms of engagement, even/especially for niche topics/interests. But network effect is smaller.

X is a total attention quagmire! At least 10 times while I've been working on this article, a brief detour there has turned into a time sink since the "For you" is so finely tuned to get you to react (even on a new account). And I think that makes it even more imperative to consider automating social media. A quick trip to a different platform isn't necessarily just a 2 minute (or whatever) task when these platforms aren't designed to respect your time or attention.

X in general strikes me as being very "manhattanized." It's where all the big players still are, and they are as visible as ever. Meanwhile, the average person on there is exposed to so many other different other people that nobody can really afford to care that much about each other. It seems like a digital subway train or a bus through Times Square.

Bluesky and Mastodon are technically and philosophically friendlier to the kind of integrations and automation that I'm calling for in this article, but they are smaller communities with lower viral ceiling. Even if you like Mastodon, it seems like a risky/isolated basket to put all your eggs in.

I think all of this creates the need to target multiple platforms to try to find the best ROI and optimal social graphs. But it all takes more time than is humanly available. So you need strategies.

Pillar 1: Analysis

Given that there is a time/attention cost to posting on social media and the results can be variable and skill-dependent, it makes sense that we should try to get as much feedback into what works and what doesn't. This is one place where APIs can come into play.

With programming languages adapted for rapid data exploration (like R and Python w/ pandas), you can load your data from social media platforms and try to seek out answers to your natural questions. What topics perform best/worse? Are certain types of audiences "chattier" or "quieter" than others? What sentiments are implied in the language, and does this impact engagement?

These are the kinds of questions I love answering with the R statistical language. If this were 2015, it would probably be kind of ridiculous advice for me to propose "learn R" in this sort of context. But it is a free tool, and AI can help you get started nowadays. To work with data from Bluesky, you can use the 'atrrr' R package. If you wanted to learn/practice this in a context relevant to the music industry, you could pull in data from Metallica's Bluesky account and apply techniques/code from this awesome free book about analyzing text data. (I might be getting a little ahead of myself here, but want to give eager readers real actionable pathways if this resonates). Since they are a massive, successful band with resources to put into their social media ops, it's worth seeing what can be learned from their approach and copied.

I used to scrape the living daylights out of Twitter with their old API, but the rules and limits have totally changed since becoming X. Post analytics are now a product/service they try to sell you. I like BlueSky because their 'AT Protocol' is friendly to the kinds of data requests we are interested in for researching/validating our own activity. For my main personal account there, I post about a lot of different topics that are important to me professionally and personally. When I pull my post data into R and calculate grouped summaries for engagement data, it is easy to see patterns across my posting patterns

At a high level, when I post about the programming environments/frameworks/libraries I use in my real work, it is well-received. In the scheme of things, I am a nobody (270 followers as I write this but started the period from zero), and the R and Laravel communities are loaded with PhD researchers and senior engineers, respectively. Those two communities engage differently, but in both cases it seems like I got exposure and people got value out of me posting. And at the same time, posting about Formula 1 and the NBA never really yielded me more than a like or two, even when I spent a lot of time and energy writing up unique analyses/models for those sports. Such content was largely being published out to no reachable audience.

If my numbers seem tiny, keep in mind my point about platform differences but also the idea of reaching the right people. Without naming anyone specific, there have been multiple times where some of the researchers who inspired me to learn programming have liked my posts! To me, that seems like a huge win in terms of validating targeting machinery/algorithms. I didn't go out trying to get anyone specific to like my posts, but it shows that social graphs and hashtags got my posts to a lot of very relevant people. Maybe your music isn't going to be the biggest new movement in the world, but what if there is a small, specific community waiting to hear it?

When we need to make money, why release something into a setup where there are few or no reachable fans/consumers? It sounds stupid, yet countless bands did it in the 90s. Some bands, like Bon Jovi, adapted as glam went out of style, but plenty of others lit massive sums of money on fire releasing hair metal albums into a market that had no real appetite for it anymore. They couldn't see that interaction on the #HairMetal tag was collapsing so rapidly.

And since I already tried to send you down the DIY route, here is the closest thing resembling a sales pitch in this article: if you hire a pro to help you with social media analysis, that is not an ongoing cost. You do not need to constantly pay someone to sit there and analyze your data and prepare the next report. If you were to engage with me, for example, the whole point would be to create a reusable script/app/pipeline that lets you stay up to date on the current state of things without needing to redo any work. If that sounds exciting to you, please contact us or reach out to me personally on social media. And if someone wants a bunch of money on a recurring basis to create monthly spreadsheet reports about your social media, they are essentially scamming you.

This is a fairly broad area, and I will surely be posting more articles, tutorials, and tools for this kind of stuff in the future. For now, I hope I have introduced you to the idea of trying to own the data from your social media activity and being able to surface insight from it.

Pillar 2: Automation

Social media APIs can also be used to create automated posts to social media. Exact strategies here vary, but the idea is to create surface area for people to find you while minimizing additional time costs.

I think there is a fine line here. It is easy to be slightly too spammy and become background noise. But part of the point of this whole article is that you are going to get buried, both in algorithms and people's minds, if you have a great song and only post about it once.

I think the key is to look for areas where media is used repeatedly on multiple platforms. There is a lot of content on social media that is basically just pictures of guitars. That seems like an easy thing to automate. If we can get a bunch of guitar pictures, on a server, we can programmatically schedule posts on various platforms (which will be executed through their respective APIs). I've seen a few programmers that post videos everyday with them coding and some kind of "Day x of coding until I become a Amazon engineer" message. I suspect the creation and posting of such content is often automated. After all, that's what a serious software engineer would do, right?

Renting a basic server from a cloud service is cheap; it's basically buying an extra pack of guitar strings a month. Although we should always be wary of subscriptions and subscription-like expenses in life, in this case I think it's basically noise if it buys back any amount of your time while expanding your reach. The benefits compound.

Automation + Analysis combined is where the fun really starts. Having automation helps give you more data to analyze! And that data can improve automation strategies.

One thing automation definitely doesn't mean for you starting out is using LLMs to post independently. Just using the APIs with some sort of "dumb" software to schedule and coordinate posts/content is a first baby step. If you haven't tried automating your social media presence without LLMs yet, what makes you think you are going to be able to handle all the complexity and nondeterminism that AI introduces? And without good analytical capabilities, you don't really know how the AI is representing you at scale (which means there isn't a feedback loop for improvement either). If AI automation of social media is your end goal, please don't start there. It's a perfect example of why companies shouldn't view AI as some sort of panacea without being willing to invest in the proper data connections to get real value. Data connections are where the value is.

Look for content re-use opportunities and tech that facilitates it

Being able to remix your content into different formats can open up totally new people for your potential/accessible audience. For example, a lot of programming youtubers will take their coding livestreams (which have a real but niche, limited audience) and turn short clips with advice or opinions into short videos. There are a few creators I follow who would not reach me without doing this because I am always sleeping or coding when they are streaming. Maybe they are leveraging automation?

If you livestream some kind of musical performance, I think the same audience dynamic might apply to you too. There are plenty of people in the world that might be fascinated by your music. If you could make short video clips out of your solos from stream, for example, you get to double dip on your effort/performance in a good way.

I find myself doing this a lot, sometimes by accident. The nature of most social platforms is to try to force you into bite size content. Not every tweet can explain everything or even land with people. That is ok. Sometimes, a post sets up a quote tweet that will land with people (because some stuff needs more characters of context to actually mean anything) or a blog post. Hey, you've probably noticed me doing it!

And when you identify opportunities to do this, look for technology that automate it or otherwise remove friction. There are always time/attention costs with social media unless it is fully automated (and even that has a setup cost). This is actually another area where social media platforms offer APIs:

Post content about real music you are writing/recording

Maybe you already do this, and if so, great! It doesn't seem like everyone is doing this (or prioritizing it), and "the algorithm" and just people's attention spans won't always reward it. But I think you have to try. Here is my shot:

It's a serious question, by the way. And choosing to launch this site has not helped things. But that is life, and I am very happy and lucky that I have multiple projects that matter to me.

Take back social media agency

If you are a professional musician, I think your job nowadays doubtlessly includes some form of social media posting to let your art participate in the attention/algorithm economy. Your career demands a professionalized social media presence, but that doesn't call for 24/7 responsiveness and nonstop video/image editing. It calls for a systematized approach.

Any sort of social media or "content creation" has the potential to remove time from writing and recording music. Even this blog is sort of a cautionary tale and a balancing act in that regard; surely in the time I spent building this site and writing articles here already, I could have recorded an entire metal album by now. But life is hardly ever that linear and absolute in terms of how we get to choose to use time, creativity, and attention.

When evaluating which platforms to invest in the most, don't neglect to look into APIs and the developer environment around the site. If a platform literally doesn't let you (or a freelance developer you hire) build tools to help your career/business, I think it becomes brutally clear who the platform is designed to benefit. And it isn't you. Image buying an amp that is welded shut in so many obnoxious ways that you can never change tubes or take it to a tech. Would that make any sense? If you are trying to build a business/career, doesn't it make sense to try to build your business on platforms that actually let you build?

Social media can very easily be addicting. It's largely designed to optimize for that, actually. It can also be a gamble of time and attention. I am trying to see it as infrastructure that connects me to others and use it accordingly. Maybe that idea resonates with you too. I think it goes with the terrain of being an artist.

I think social visibility is something that can be engineered and is worth trying to engineer. Once we are viewing things from an engineering lens, naturally we have to define what we are optimizing for. You can go for all-out maximum reach with slop content (because any content that is any more specific than needed or takes risks will turn some people off) or you can try to the users that might be open to content that is uniquely you.

What kind of "automated glue" would be useful in your social media presence?

If you found this article interesting or helpful, I humbly invite you to follow us on Bluesky or X to stay informed about other content about the overlap between the tech and music industries.