My POV on using Spotify’s ad space

This is a pretty “finger in the wind” style look at it because I have pretty limited experience in ad-buying. That said, I think there’s some common sense in the following.

Spotify makes very little off advertising. I saw an article somewhere that specified how much exactly that figure is but can’t find it right now. Failing that here’s some data that lets us triangulate it:

For the larger streaming industry this is how the profit break down looks according to pando:


Of that $1.4 billion, almost half came from customers who pay to use Spotify, Rdio, or Pandora, amounting to $628 million. A slightly smaller chunk, $590 million, came in the form of performance royalties which are collected and distributed by the performance rights organization SoundExchange. (The vast majority of these come from Pandora; services like Spotify and Rdio are “on-demand,” not radio, meaning they negotiate directly with labels). The remaining $220 million came from advertising-supported streaming and was generated in large part by YouTube and Spotify’s non-paying customers.

So on demand streaming is already the smallest profit chunk (let’s ignore Sound exchange distributions because from what I can tell that represents an outflow of cash for Spotify).

On top of that reports

(This is their assumption):

Spotify makes 5 percent as much advertising to its 15 million non-paying frequent users as it does from cash from its five million paying customers. According to Pandora’s recently-released financials, we estimate it makes $5 per year from non-paying users, so that seems like a good number to use for Spotify too.

So you can see that as far as the business model goes it’s all about subscribers for Spotify. What this means is that I imagine management is prioritising subscriptions over ads and probably not trying too hard to give advertisers bang for their buck.

I noticed that some of the more recent ads on Spotify say things along the lines of “If you buy X brand you’ll save enough to subscribe to Spotify”. Which might seem like a clever line to a copywriter in an agency that doesn’t understand Spotify’s business model, but to me it sounds like Spotify’s marketing team is even using the actual ad spots to drive subscriptions.

Advertising appears to be being pushed out of the mix in favour of subscriptions, to the point that it will be used as the stick that drives subscribers, the carrot being ad-free listening.

So that to me makes ads on Spotify a relatively caustic proposition. I’m sure there are some uses for it (penetration?)  but I think it will become increasingly less useful to place ads there. Spotify may start requiring that you include subscription CTAs or overcharge to satisfy increasingly steep KPIs that demand advertising keep up with subscriptions (an impossible task but businesses don’t seem to let logic get in the way of their KPIs).

Radio actually seems like a smarter choice because at least it’s consumed with the expectation that there will be ads and talking heads. That gives advertisers more room to be memorable with clever ads or jingles.

I think podcasts are a very underutilised advertising medium. Reasons they are good: The ads tend to use a code-based model which allows you to measure impact and track ROI. A good host can read advertising in a way that makes it bearable, sometimes even hilarious. A lot of online businesses like Squarespace and Audible seem to be jumping onto podcasts where as I am seeing more and more FMCGs and traditional brands using Spotify. That to me indicates that I’m onto something because the former are usually more onto it.

I wish I could spend more time to investigate my hunches about this stuff but I have a day job and hobbies. Not to mention I’m lazy!

Engagement rate isn't engagement rate! Finding the right reporting metrics

I’m currently struggling a little bit with different social metrics. There’s no consensus out there on what is best.

Some say Facebook engagement rate should be measured as (likes+comments+shares)/(page fans) (often referred as page engagement rate) others say it should be measured as (likes+comments+shares)/(page reach) aka post engagement rate. EDIT: Actually I don't know what that approach is called but Post engagement rate is something different again. 

Personally I think post engagement rate is better. It doesn’t get skewed by the number of posts (page engagement rates go up the more posts you have because it doesn’t factor post numbers in. You can then divide it over number of posts but it just starts getting weird). Also I don’t think that fan numbers are as important anymore now on Facebook because Facebook is dialing down organic reach so dramatically anyway.

But then you also need to think about whether or not you want to includes clicks in engagements. Tricky!

Then on top of that when you have actually decided on how to measure your engagement rate you then have the issue of communicating it to a client. Because clients usually think there’s only one way of doing it. Then if you've chosen a better but less popular measuring system, you will have a hard time finding industry benchmarks. Total clusterfuck.

Taking a step back

I think the underlying issue is also that we are only just starting to get to the point where we say “Hold on why are we evening measuring any form of engagement in the first place?”. Facebook had a lot of people believing that engagements were a great proxie for share of voice and a great way to get people to buy things. They had it packaged up as “stories” the more “brand stories” your users are telling the more likely their friends are to buy your stuff.

Well that seems a bit shaky now and I do sort of role my eyes at the term “stories” when it pops up on my spreadsheet. Beowulf… The Godfather, those are stories. Clicking like on a brand post… Not a story.

I think engagements are still a nice proxy for excitement. If something excites or arouses people they will probably like and share it. That then implies we've created impactful communications. Which in turn are more likely to stick in peoples’ minds. Add reach to that high quality content and you have a good basis.

An experiment I’d like to do

I’d still like to see how clicks to site and engagement correlate. Or any CTAs and engagement for that matter. Just plot them all out on an X and Y axis and do a regression analysis.