Hacker News Reader

Perhaps I’m naive but my view is that online advertising should have targeting no better than bus stop ads (coarse geographic) and conversion should ideally not even be measurable. When I buy shoes I prefer to not even say which campaign made me do it (just like the bus stop ad).

I find the whole jargon of this post dystopian.

I don't know. It isn't that I think targeted ads are bad. I think it is simply jarring that humans are that quantifiable, and separately, that I am skeptical of the privacy policies of the gatherers of this data.

My Instagram feed shows me lots of startup clothing companies for young to middle aged men, and I finally tried some, and damnit if they aren't decent jeans.

If we accept that ads are not inherently evil, and we accept that ads (when not manipulative or selling unethical/addictive products) are more beneficial to a recipient if they're relevant, then targeting (if done ethically and stored as anonymously as possible) is a great thing.

I find that the ads are actually more straightforward than traditional ads. They literally say "These jeans are stylish, durable and American-made". I didn't sit through 30 seconds of Brett Favre throwing a football to signal that I'm just like him.

I think they are bad if they use context I didn’t want to provide. I’d be happy to give any website a pretty detailed profile of interests and “cohort” if I knew that’s where it stopped.

But I don’t think ads can target ethically and with respect for privacy. I’d love to be proven wrong.

I'm also not opposed to targeting in principle, and in fact I'd love to reach an end state where a company - whether you call it an ad company or otherwise - does 98% of my shopping for me. But I feel like we're in an uncanny valley in which adtechs are hoovering up more data than they'd even need to do that, but with no incentive to push consumers toward products we'll actually be happy with. Facebook is probably better off if your new jeans wear out in a few months, and Amazon can't even be trusted to reliably ship you the products they think they're shipping you, let alone vouch for their quality.

I don't like this assumption that we all have piles of disposable cash (or credit) - our main concern is the struggle to find the right "fit" for using it and therefore targeted ads are improving our lives. Due to the size of the global economy, there are many more products that fit you than you could ever buy.

It can only be considered "a great thing" as a response to an earlier status quo, but it's not one in and of itself. Taken to its logical conclusion, in a world where you can encounter dozens of targeted ads in a day, what happens when every single one you is a fantastic fit for you? It's surely not "a great thing" for mental health to have things you truly desire shown to you at a frequency faster than you can afford them, especially in a world where credit loophole agencies like Klarna prey on the vulnerable. And this isn't even opening the data privacy can of worms which is required to achieve that, or the fact that these platforms are competitive and reward the highest spender for any given perfect fit.

At least let's try to stave off the dystopian future for a little while longer and hold on to the notion that consumerism doesn't have to be optimised into ever more efficient workflows. Even if you enjoy shopping, doesn't this take out some of the excitement and satisfaction? How long now before some unknown third party is monitoring your spending habits to deduce your payday, sells that information to another third party that buys data from a smart speaker company which overheard you discussing jeans in a public place and automatically sends a message selling the lead to the highest bidding jeans shop who has the product in stock in the waist size which is available from another this party and your virtual assistant puts an event in your calendar telling you to be at home to receive it?

I'm guessing you've never tried to start an Internet-based business. Being able to target ads is often your lifeblood of your company. Or have you tried and somehow found a different method?

I run a company where about 70% of my audience (in 2018 no less, so likely a lot higher now) use adblockers. To me, any amount of search ad spend or similar is pissing money down the drain.

For reference, our marketing strategy is opt-in email newsletters, organic social media, the occasional sponsorship of events, and natural word-of-mouth (no incentive). And we've been in business for 12+ years.

Search ads are very effective and are paid on a per-click basis, so you really can't go wrong with them.

>opt-in email newsletters,

Opt-in, or "kinda opt-in"? Or you just don't know and trust whoever publish that newsletters?

Seriously newsletters are the most annoying ads I get and by far. So many are not really opt-in, but instead use dark patterns to make you sign up.

Fully informed and "Have to click the checkbox" opt-in.

I absolutely have not. But I don’t think the possibility of buying tracking ads is somehow excused because e.g “you couldn’t start internet businesses otherwise” or even “most of the free content online would disappear”.

Targeting doesn’t have to imply tracking or trading in my info. Facebook knows my age and gender, and some of my interests (from groups etc). That should be plenty. That means a company that sells kitchens can’t target people “interested in kitchens” in the sense they have recently searches for it or visited a site of a kitchen manufacturer that told Facebook. They can target people in country X of a certain age. If that doesn’t sell enough kitchens then just close down me kitchen business or buy a bus stop ad but don’t follow me around the internet.

Note that plenty of advertising like bus stop ads, especially if there's any QR code, is often tracked for conversion (by having a unique reference code embedded in the URL in the QR code).

That would make it much harder for me.

I run a theater troupe. We do a little bit of Facebook boosting. Our customers are spread thin: there are people who want to see Shakespeare, but they don't cluster in one place.

We do use geography in our buying; nobody is going to travel 100 miles to see us. But if I showed it to 1,000 random people even in a 20 mile radius, the odds are that 999 will be completely disinterested, and the 1 remaining person will be doing something else that weekend.

On the other hand if I show it to 1,000 people who have shown at least a little interest in theater, or Shakespeare, or literature -- and who have the money to spend $20 on a theater ticket[1] -- maybe I'll find a few people who would actually come.

I wouldn't show my ads on a bus stop. It would cost a fortune and probably do no good. I spend $30 or so, and I get a few extra bodies out, because my ads are targeted.

This is just an example I happen to be familiar with. Perhaps it's a little too abstruse to be applicable. But it's what I think of whenever I hear people railing about all ads. There are a lot of things for which blanket advertising just wouldn't work.

I'm not even trying to turn a profit or pay salaries. I just want to do some good art that people might enjoy. It's worth getting the word out for it, and not pester every single person on the planet about it.

----- [1] I make it as cheap as I can, but theater space costs money.

But when you show it for 1000 well targeted people, what are you paying?

The key to all this is that it’s expensive because they are targeted. If you could show 1M less targeted people for the same money you would show it to the same people.

Sure, 999/1000 people would see an ad they weren’t interested in, but that’s inevitable with less intrusive tracking.

I also wouldn’t mind being shown a very well targeted theater ad if I had volunteered that info. But if that info is gathered without my consent I just shouldn’t see the ad, and that’s regardless of whether 90% of the free content of the web and a huge number of small businesses die in the process.

About thirty bucks on Facebook. I couldn't afford a TV ad or billboard or even a bus stop sign.

> You can simply control by making only one single change segregated to a control and treatment geo/group before and after, measure delta.

Because nothing else is going on in the world besides your ad campaign, right?

Like you I find the comments about the ad industry here on HN largely ignorant, but I also think most advertisers and agencies are dumb as rocks, or at best high on their own supply. Like, a piece of analysis I saw recently had some claimed weirdass causation along the lines of, X days is the optimum period between first indication of intent in a product and showing an ad for that product, if the user didn't convert immediately. They had run an experiment like you suggest with varying intervals for audiences. But, the way better correlation was with stimulus payouts. That wasn't in their model, because their model only had the inputs they controlled, before and after, and measured delta.

Everyone is selling snake oil because every advertiser is demanding snake oil. It's so fucked up.

Yeah definitely 'multi-touch attribution' ad tech is a lot of crap a lot of the time especially for big brands stuff. And lol got to love advertisers selling advertising it's like one of the launch websites you see on Show HN sometimes that just have a bunch of #words that have no meaning ;)

it's the flat out 'ads don't work you can't know they worked' stuff which bugs me that is evidently hard to simply explain

As someone who doesn't work in advertising but is insterested in the space, I find your comment interesting but it doesn't help you are not providing much data for your hot take either. You mention Kickstarter, but looking at the list of crowdfunded projects, most of the successful ones were quite a while ago, if anything the most recent ones looks like scammy crypto coins, which would suggest advertising works rather for bad projects, a net negative.

https://en.m.wikipedia.org/wiki/List_of_highest-funded_crowd...

> You mention Kickstarter, but looking at the list of crowdfunded projects, most of the successful ones were quite a while ago

If you want to see Kickstarter projects that failed because of lack of advertisement, I suggest you to go Kickstarter itself[1]... you may have better luck than going on the List of highest-funded crowdfunding projects, which is mostly unrelated to Kickstarter.

You'll see that most of them doesn't require much money to works. The only reason they doesn't get that money is their lack of advertising. Sure they could buy ads, but you'll see that they all are pretty niche too, they can't just get the same advertising as a bus stop, that would be just too absurd and cost them 100x more than their current goal.

You can believe they just doesn't deserve to exist, that's fair, but that's not the point. The point is to show the importance of advertising and how niche project just couldn't reach their audience if you make it harder and harder to reach.

[1] https://www.kickstarter.com/discover/advanced?raised=0&sort=...

Advertising is cancer and it’s a good thing when less money is spent on it and less of it is thrown in peoples’ faces.

The ideal system is one where people pay for services and see no ads on the services they pay for and nobody can monetize their data.

I can't afford to pay for all the stuff I use. FB, IG, YT, etc would cost a fortune. I'll happily watch my ads, thanks.

I’m not deep in this space but was thinking something along the lines of your response. If I understand it right, advertisers will lose efficiency here but I’m skeptical that it’ll cost ad networks much because advertisers will still need to reach an audience but will now suddenly be less effective at targeting.

If Facebook ads are so good, why is it that I basically only get shitty, general ads?

I am a single geek with disposable income, why is my feed not filled with ads for cool tech products? I just checked and the only Facebook ad was for cancer awareness.

Are you actually allowing the tracking?

I am an Android user. I don't have a choice :(

You do. Install blockada. No more ads and trackers, even in apps.

https://blokada.org/

You have the choice to both adaway and firefox with the usual addons. You can also choose to not have google on your phone.

Do you have a PiHole?

in my experience, fb ads are a mixed bag. they tend to give you the illusion of control without the commensurate results, converting at a noticeably lower rate. sometimes prices are low enough to compensate for the lower conversion, but not usually.

though it might be changing with the normalization of the fb marketplace, folks aren't usually looking for the things that require advertising inducement on facebook, so there's a significant intent/context mismatch.

I come from the user side and I am extremely pleased to hear this news. Advertising on my device has gotten completely out of hand, I never asked for it and I don't want it.

It seems like people have already forgotten. When TV ads gradually got less and less effective in the 90s. What was the result?

Incredibly intrusive constant bloody annoying ads. Not less of them, far, FAR more. To the point the TV became unwatchable.

>Good or bad or neutral, FB's targeting specifically works amazingly well and better than anything else for many buyers. Especially for direct response ads.

A recent study says otherwise.

https://pubsonline.informs.org/doi/pdf/10.1287/mksc.2019.118...

Do you have any studies supporting your claim that it "works amazingly well"?

I buy political ads for a living. I can (usually) make an immediate positive return on investment running ads for donations or growing an email list. Been doing this for years and has raised millions for our clients.

I have tested and spent money on other platforms none of them work nearly as well for our clients. Search ads work but I believe that's far more of what OP is talking about. If your website is already #1 spot and someone searches to donate they would probably do it already getting there quickly without the ad. The problem is when others or even the same party parent org buys that top spot to get that traffic.

As someone not familiar with the advertising world, it wouldn't strike me as particularly surprising to learn that political advertising on fb works better than other domains. Not only do people post entire novels on their political opinions, but they're constantly exposed to debates and likely are more likely to feel that the ads are important to react to in the moment.

I think it's partly the 'housewives' older demographic the people who spend a lot of time on the platform sharing political stuff they read.

But mostly it's the data and attention on higher quality ad space. We upload donor data and target that + lookalikes.

We can't do that with google anymore, using 3p data on a DSP is both way worse quality data but also wayyyy worse quality ads. twitter never worked for us either.

A simple no would have sufficed.

No one is arguing that data brokers don't make millions selling ads. The argument verified by the study is that demographic and location segmentation doesn't work when targeting ads.

Demographic and location segmentation is just a shitty way to target and is a prime example of looking for your keys under the lamppost because that's where the light is. That said, your point is unfortunately very relevant because many advertisers still do this.

> A simple no would have sufficed.

Actually, I think it is pretty important that they explained that they are an expert in the field, and that you, a not expert, don't really know anything about the space, and thus your opinion is quite literally, not as informed as theirs.

It's a real study with actual data.

First, I'm talking about better targeting than demo + geo. That's what 'old' media has had for like a century. What works for many D2C and DR advertisers specifically on FB as an example is a combo of: retargeting, customer list upload (using lifetime value e.g. someone who bought many times in past sees more ads), lookalike from those lists.

probably wasted a bunch of time but I went ahead and searched some things if you care to learn more.

There are a ton of sources showing value of this data and ROI. FB just recently posted this study they did with Group M which is like the biggest media buying conglomerate they touch everything. FB's research is pretty good though obviously they wouldn't publish something totally negative.

https://www.facebook.com/business/news/insights/a-full-funne...

Another recent piece surveying marketers showing IG stories have better ROI than Snap (that's our experience by a long shot too)

https://archive.is/U0NDu

App Install ads are another huge vertical for direct response which drives a ton of revenue.

I just looked at Zynga 10-k. quick word search they say from FB, google ads "In 2020, 2019 and 2018 we incurred $583.1 million, $377.2 million and $157.7 million, respectively, of these player acquisition costs."

I don't see quickly if they report a direct roi (im sure that would be closely held..) or cost per active user acquired but they do mention a ton how it's revenue driver and list all the risks from iOS privacy changes, and mention risk with more volatile ad market with covid.

and privacy/data rules impact Zynga double because they have to pay FB for instance for new users, but then they sell ads in their games.

They specifically talk a lot about building up their own 1P data (like FB) to up value on their inventory (or evidently own ad network if they are buying/building one). But the risk is iOS will be blocking a lot of this, and in my experience 'open market' ad exchange 3p data is not even close to as good as FB. and see link about build/buy their own ad network

https://investor.zynga.com/static-files/c0a44588-72cb-4e6e-9... https://investor.zynga.com/static-files/5b5771d0-52d1-4b91-8... https://www.adexchanger.com/ad-exchange-news/zynga-is-buildi...

No offense I don't think you know more than they do about their own businesses (and like also questioning mine...).

It bugs me that people who don't seem to work in the industry just assume they know better and that these million dollar revenue companies (sometimes billion $) are just throwing their money away because they are what. Stupid?...

Like Amazon is another great example of the high value of the type of data apple wants to get rid of. Not many companies will pay $20 for a $10 purchase at any scale. Maybe for Dollar Shave Club type business they do since they have recurring revenue projected.

I only found one kind of hit on scholar there aren't any direct response articles easily findable.

I can't read this one but the title says "Personalized ads on Facebook: An effective marketing tool for online marketers"

https://www.sciencedirect.com/science/article/abs/pii/S09696...

and also kind of funny here is one showing FB targeting worked really well for clinical trial recruitment

https://www.researchprotocols.org/2018/1/e20

None of those links disprove the arguments in the paper.

No one is arguing that using large platforms vs not using large platforms gets you a larger pool to work with.

The argument is the actual segmentation they champion is useless.

I will be happy to be corrected if I'm wrong, but I have a hunch that your clients skew conservative. Not based on anything you've said, just based on your apparent success on Facebook as a platform. Would you be willing to confirm or deny that?

That study isn't about Facebook targeting, right? It's anecdotal, but FB/Instagram ads are noticeably better for me than anywhere else: they're probably the only ads that have ever resulted in me buying the advertised product.

It blinds all the vendors and one of the panel data sources was Facebook.