Marketing Monday Written by Jon for June 8th

You can't please every algorithm.

There was a bit of a craze a couple of years ago toward front-loading Meta ad budgets in the early days of a release¹. The idea is sound and simple: Spotify's algorithm rewards early activity on a track. So if you put most of your budget toward the first few days after release, the reasoning goes, you'll reap the rewards in algorithmic streams for months and years afterward.

To make it concrete: Let's say you've got a budget of $500. You might plan to spend $100 in ads on the first day, $300 over the rest of the first week, and the final $100 over the rest of the month.

Done right, it works – really well, actually. But it can also go really, really wrong.

The problem is that music streaming ads need to work with two algorithms.

If you're running Meta ads to drive streams on Spotify, then obviously it's not just a matter of pleasing Spotify; it's also a matter of pleasing Meta. But the tricky thing is that, while Spotify's algorithm cares about release recency, Meta's algorithm doesn't, and in fact Meta ads nearly always take time to reach their highest level of performance. They literally have a Learning Phase during which Meta is testing where to show ads, and my own data shows that, on average, campaigns take 47 days to reach their most efficient level of performance.

You can see the problem.

If you spend all of your money in the first week after release, you'll run the risk of wasting it on an unoptimized Meta ad campaign. But if you wait to spend most of your money until you've dialed things in on Meta, you'll run the risk of missing the critical release window on Spotify.

Thankfully, the solution is simple: Just have unlimited money. That way you can spend a ton upfront and also spend a ton later on when the campaign becomes fully optimized. Win-win.

Jokes aside...

Here's my real two cents on the front-loading strategy.

It works super well if these three things are true:

1. You regularly release the same sort of music (down-the-middle EDM tracks, for example)
2. You've trained your ad account and pixel to know your audience
3. You know what kind of ad content works well for your music

Basically, if you've got your release strategy boiled down to a repeatable system, then yeah, front-loading your budget can be super effective. Meta will already know who to show your ads to, so you won't really be running the risk of pumping a bunch of money into a poorly optimized campaign, and you'll jumpstart Spotify algorithmic traction immediately following your release.

But if any of the three above things aren't true – if you're releasing songs in different genres, or if your ad account isn't well-trained, or if you haven't yet figured out what sort of ad content works for your music – then you will almost definitely waste your money by front-loading your ads budget.

I wanted to use this specific tactic as the entry point to a bigger-picture discussion of algorithms.

But (unsurprisingly) I'm finding myself crunched on time, so I'll leave you with the following quick take:

When you use a platform, it's wise to familiarize yourself with the whims of its algorithms. But it's even wiser to remember that the whims of any platform's algorithms are only proxies for the whims of real people. You can't please every algorithm, and anyway it wouldn't matter if you did. So don't get caught up in the minutiae of perfect formats and secret hacks and whatever else us marketers are on about.

I think only a vanishingly small number of your artistic decisions should be dictated by the demands of algorithms. I think just about everything you do should serve the interests of your people.

Good luck.

– Jon

Jon Anderson Founder @ Two Story Media Surprisingly Bad @ Being Unbiased

🫆 100% Human Guarantee I wrote this, not AI. Today's random photo: Tom's mentioned Dorothy Sayers like 80 times, so I finally got around to reading one of her books last week. My gosh the ending was satisfying, I absolutely love how that guy died.

P.S. – Last week, I wrote about another popular piece of music marketing advice: releasing a song every 4-6 weeks. You can read it here.

P.P.S. – Tomorrow at 4pm EST, I'll be doing a live Q&A inside of my marketing membership. (I do these each month, but I hardly ever remember to mention them in my normal newsletters.) If you'd like to join the chat, join the Club.

¹This approach was popularized by Jend. I'd like to note that Jend a) is generally great at ads, and b) does not advocate for front-loading as a one-size-fits-all approach. Here's an interview I did with him last year where we talked through some of the nuances.