Hacker News Reader

Comment by Ozzie_osman | Hacker News Reader

The way to measure whether the data-driven ads are a scam is not by saying "this ad is irrelevant to me". It's by the advertiser looking at their ROAS (return on ad spend). If the algorithm is optimizing for showing you an ad, there's a pretty fair chance that the advertiser has done that math and decided that in a world of imperfect data and algorithms, you approximate enough the type of person who is worth showing an ad to... statistically.

In fact, getting to a world where ads are perfectly targeted would be one hell of a privacy nightmare.

View on HN

> there's a pretty fair chance that the advertiser has done that math

As an occasional admittedly naive experimenter with ads – it’s far more likely that the platform has done the math and showing you the ad is profitable to them, not the advertiser.

Google et al love nothing more than to burn your entire ad budget on cheap clicks from countries that will never buy anything. But the number of “ad conversions” looks high and graph go up!

Getting these platforms to do what you want is a literal full-time job. Hell, in bigger companies you have entire teams doing nothing more than bending over backwards to get these platforms to do the useful thing instead of setting money on fire. The algorithms love setting money on fire. It’s a lot like giving your 3 wishes to a genie, you have to be very specific.

I’m not sure this is true. Facebook makes a lot more money from me if my ads are converting, and I continue to spend money day after day. If they just burn my money and I’m not seeing enough conversions out of it, then I kill my ads and start trying my luck with Twitter or Pinterest ads..

> I’m not sure this is true

I can assure you it is very true that it takes a lot of skill to get ad platforms to do what you want. They do not care about you. The algorithm will make The Metric go up and that’s a threat. It’s your job to make it a good metric.

While people relying on a 3rd party platform are under the whims and methodologies that an opaque 3rd party platform may provide, in first party AdServing (which I am involved with), the algorithm is not tailored to "your" metric. Data driven analysis is done by all major online retailers. It's quite effective and part of the reason that features and products seem to disappear, more often than supply chain disruption would indicate or the common perception might belie.

While I'm sure some companies have some "reliable" 3rd party + 1st party metrics that affect their bottom line, they don't need more until the trends start to invert, and then they just don't have the data for marginal short-term experiments. This causes these companies to make sweeping changes that a 3rd party will not help tool around...leaving the company to decide the changes either 1) help climb returns (overall), 2) hurt returns, 3) most commonly, show nothing at all. It's flailing in the dark, hoping for an external force to turn things around. The company that is still relying on someone else's platform are dinosaurs and small players. Sometimes they are able to climb out of it due to externalities or developing their own platform.

Perhaps they optimize enough to give you the minimum conversions to stay

> Facebook makes a lot more money from me if my ads are converting, and I continue to spend money day after day.

There are a lot of small companies in the world. Would Facebook miss you as an advertiser?

I don’t think so as only the advertiser has data you need.

Let’s say sales increase but your advertiser just burns 20% of your spend because it can and you won’t know. That’s the scam part.

It’s not that your spend doesn’t increase sales. It’s that part of your spend is wasted and you could increase sales more if it wasn’t scammed away as wastage.

I just bought a dozen bagels for $20. I’m happy, pretty good deal. But if the store was throwing away a dozen bagels for every one I bought, that’s wastage. Just because I’m happy doesn’t mean it’s not a scam.

Reply on news.ycombinator.com