I love the fact that I get internet content for nothing as it's paid for by big advertisers who are to stupid to realise that I never purchase anything that they advertise.
Yes and no. Conventional wisdom inside ad tech is that fraud rates are about 30% on average. Typically lower for video. So there's real value in there (in terms of ROI for the advertiser) but there's effectively an entire margin of fraud where advertisers are bearing much of that cost.
Video ads are intolerable. Whenever they exist, they spawn an arms war of dark patterns to get and stop them from auto-playing, without warning, at maximum volume. The Facebook Android app is gawdawful for this; I just discovered that you can put it into some sort of a data-saving mode, which seems to stop auto-play for the moment... but time will tell. And though we're supposed to be bagging on Facebook here, YouTube is really bad with some of their in-the-middle of content ad spots that I've been noticing lately. I'll be watching a peaceful little Timeline documentary, and right in the middle of an interview the TV starts blaring pop music for something I am never going to buy. Especially galling when it's something produced for cable TV, so it has commercial breaks built into it naturally, where the ads could be inserted unobtrusively.
I think what’s shocking is how awful a lot of the targeting is. Even before I started getting super aggressive about blocking all trackers, I regularly got fed ads for products and services I didn’t need, and in languages I couldn’t understand.
The most confusing is that I keep getting ads from the same company I bought my wedding ring at. Like, I’m not divorced, and I don’t need a second. Why would they re-target me?
> The most confusing is that I keep getting ads from the same company I bought my wedding ring at. Like, I’m not divorced, and I don’t need a second. Why would they re-target me?
I'm guessing there's a good reason for this. Did you visit the company's website? And did you buy in person? If so, you're seeing the online to offline problem: specifically, it's hard to correlate your online data with your offline data, so it's hard to tell the difference between someone who already bought a ring from somebody who hasn't yet bought a ring. And I'd bet most people who visited the website haven't bought a ring yet.
So you bring up an interesting point: this is an externality of 1. wasted ads, 2. irritation for the user, 3. a useless invasion of your privacy.
would they want to solve this problem? It would improve their targeting but also lower their advertising base significantly.
Ive used a shopping cart product for about 6 years, I still get ads for their company on my phone. No offline to online issue there and they still dont solve it.
Do you mean the online to offline issue? It's a pretty big thing right now because it's actually two issues: (1) is being more efficient with ad buying, and (2) is knowing whether your online ads are actually getting people to come in and buy stuff.
Hypothetically, let's say current ROI is $1.10 for every $1 spent. We stop spending for people who already bought a diamond. Our ROI might go up to $1.12 for every $1 spent. That means the ability to avoid targeting people who already bought a ring is worth $0.02 for every $1 we spend. A company comes and says "for $0.01 on every dollar, we'll give you that ability". Of course you'd take it.
That's more or less what's happening now.
> I've used a shopping cart product for about 6 years, I still get ads for their company on my phone. No offline to online issue there and they still dont solve it.
Yeah you're hitting on another theme of modern day digital advertising. Our collection of data far outstrips our ability to actually act on it. So you'll commonly find situations where you'd assume they could know based on the data they probably collected, but for some reason they don't act on it.
>> Hypothetically, let's say current ROI is $1.10 for every $1 spent. We stop spending for people who already bought a diamond. Our ROI might go up to $1.12 for every $1 spent. That means the ability to avoid targeting people who already bought a ring is worth $0.02 for every $1 we spend. A company comes and says "for $0.01 on every dollar, we'll give you that ability". Of course you'd take it.
But this is only if you are looking at 1 user right? The advertiser is making the call between the incremental income from better tracking vs the extra income from more impressions.
It seems as if they are choosing the extra income from more, but worse impressions. If they are not solving the online ad to online purchase, which in my mind, they arent doing because it limits the advertising base, which limits ROI more than targeting
Weddings are a high-margin business, they can afford to spend a lot on advertising and not be too careful about their targeting. The advertising service is happy to take their money. It's not that we can't do better, it's that there isn't enough incentive to do better at any level.
I don't think it's clear that advertisers are bearing the whole cost, at least not directly. Google, Facebook, Twitter, Snap etc. all sell placements via auction, so the advertisers themselves are determining the price. Presumably average fraud rate is already "priced in".
Agreed with you there, to some extent the fraud rate is already "priced in".
What I mean is actually this: Google, Facebook, etc. are opaque about fraud, to the point of misreporting metrics in Facebook's case. Let's say hypothetically that the advertiser had an oracle for whether any individual ad view was fake. I'm pretty sure the advertiser would be paying less. That difference is what I meant by bearing much of the cost.
Google/Facebook's marginal cost for the fraudulent traffic is basically just the cost of infrastructure. The publisher (if there is one) is often the one causing the fraud, so they're not really "bearing the cost."
The fraud rate is only priced in if every member of the auction has that oracle though right? Otherwise, they are unaware of the fraud and bidding as if their isnt. Inflating the ad price
Yes and no. You have a point, but I don't think you need every bidder to have this oracle.
If only one bidder could tell the difference, that bidder might just buy fewer ads because ROI estimates will be lower, so fewer ad buys would look like positive ROI. Of course, that's under the assumption that fraud inflates performance metrics, but I think that's a pretty safe assumption.
Watching Twitter attempt to deal with the cryptocurrency spammers that chain on to every half-popular tweet leads me to believe that it really is just incompetence (+ being a hard problem in general).
For example, their latest solution is to prevent you from having the exact same name as a famous person. Notice all the bots now have the name "Elon" or "Elon Mu5k" when responding to Elon Musk.
My previous company stopped advertising on Twitter for precisely this reason. Our own analytics showed that the majority of the clicks we were getting and being charged for were obviously either bits or other spam accounts. It was actually crazy what we were being charged for.
