I Built a Botnet That Could Destroy Spotify with Fake Listens

A lot of these fake X articles miss an important point: that advertisers do not pay for clicks/listens/impressions/"eyeballs"/"likes"/[insert obvious user action that you actually, in fact, do "pay" for] but for value delivered.

As an advertiser, I will first find a way to track orders and customers from a channel, then pay for that channel. My calculation is simple: the channel brings in X amounts of Customer Lifetime Value (CLV), and I pay Y for bringing in X. If X/Y is high enough to justify Y, I keep the channel. If not, I don't. In other words, if 9 out of 10 clicks from FaceBinGle are fake, but my CLV is $1,000 and acquiring that customer cost me $10, I'm going to keep buying those 10 clicks.

In fact, what happens is that an efficient advertising market will end up with an efficient price and channels end up having the same X/Y, whatever X is (can be CLV, can be Customer Acquisition Cost, can be Cost Per Order, can be Cost Per Visit, whatever the customer base is optimizing for).

Taking this to its logical conclusion, if the number of fake users does not affect the number or activity of real users, the only person who really pays for the fake listens is Spotify, because of the royalties paid to bands which have a fixed price per listen.

So as the average price per listen that Spotify can charge advertisers drops as listens become on aggregate less valuable and the CLV models adjust accordingly, Spotify is then incentivized to find a solution. Unless they've already baked the cost of just letting the scammers be into the economics of the platform. But "listen farms" should not, in and of themselves, scare away advertisers except for the very basic ones who have not yet come upon the principle of tracking ROI.

>As much as I love the idea of having an army of robots working feverishly to bring me riches, my conscience prevents me from doing it.

I wish more people (like those involved in high frequency / algorithmic trading) would share such ethics.