Online advertising

[Sidoh]

The problem I have is that the cost isn't well known. Nobody thinks about how the advertising agencies work, or how Google and Facebook make their money. And they certainly don't directly disclose their business practices to the users. Instead, users see free software and services thinking that they're free (and they're not). That's pretty deceptive.

Sure, and it probably will never be known.

That said, I think building a product for sale makes it much more difficult to be innovative. That's exactly why I'm saying I actually like Google's business model. They make cool stuff. I don't mean to say it's "free" in the sense that they're not making money, but it is free in the sense that I'm not paying money. Companies like Google also al Maybe some day down the line there will be a day when it's obvious that the hidden costs were too high, but I'm not nearly as afraid of that as you are apparently.

I don't think their business practices are quite as opaque as you're suggesting. Some of the details are hidden, and some of those details are probably relevant to consumers, but again, I don't seem to be as concerned about those details as you are. I don't think I know less or that I've not thought about it as thoroughly. I think maybe I'm just less risk averse.

Why not directly disclose the cost to the user if it's so great? Why is it secretive? Maybe advertising agencies really suspect that users wouldn't appreciate the cost at all or maybe they think users wouldn't care.  Either way, a price tag is nice.

Because disclosing everything is probably bad for business in multiple ways, and I don't think it's at all fair to attribute all of that to justified scorn from their users. 

And by the way, Google is among many that buries its policies in lengthy legal jargon too. At least it makes users aware of policy changes.

Sometimes, I think legal jargon is probably more necessary than it seems. I think that stuff like privacy polices tend to sound pretty windy and convoluted, but a lot of that language is necessary to cover one's ass.

I can only speculate what kinds of information advertising agencies collect and how they use the data. Whether it can identify me as nslay, my Google ID, a cookie number, an IP address, or a behavior pattern is irrelevant. I don't like the idea that Google, for example, can build (and probably does) an almost complete profile of my web surfing history (since many sites host Google Ads). I personally don't want to be tracked and I am opted-in by default. I have to 1) know that I am being tracked (which is generally kept hidden), 2) Find a way to opt-out (if any).

Of course it's relevant how targeted advertising works. Consumer data used in online targeted advertising isn't stored server-side.

People tend to think it works like this:

As you browse the web, advertising companies build a profile of you. You as a person are tied to this profile. When you request an ad, the ad servers know who you are, take a peek at your profile, and decide what ads to show you based on that profile.

That's not how it works. It works like this:

The data itself gets placed in a cookie. You as a person are not identified when you're requesting an ad. Data relevant to you as a person is sent along in the request for an ad, but it's usually very boring data. More importantly, you have complete control over it. It's transient -- all you have to do is delete your cookies.

When it comes down to it, this "profile" model is generally a really bad idea from a business standpoint for multiple reasons:

1) It's far more difficult to maintain.
2) It's not nearly as flexible.
3) I think marketers would much rather use their own CRM data than some profile someone else has built. If some retailer wants to run an ad campaign targeted at some of their biggest spenders, they have that data. Why would they want to use some bullshit profile built by someone else?
4) As you mention, there are some serious privacy concerns, and despite your pessimism, it's a really important issue. It's not something advertising companies raise their nose at.

I don't mean to say that there aren't creepy ad-related things on the Internet that work by building a profile of you. I'm just saying it generally doesn't work as well as this much less scary data-in-cookies model.

If you don't want that data to be sent to advertising companies, there are quite a few things you can do to prevent it:

1) Set the DNT header.
2) Disable third-party cookies.
3) Opt out everywhere you can (it's more than you think). Services like SelectOut help by opting out of a bunch of things at once. (Amusingly, I worked with the guy who owns SelectOut. Perhaps even more amusingly, it seems to be having technical issues at the moment). Companies respect optouts. It's a PR disaster if they don't, and it's easy to test if they are.

And again, as I pointed out: Anonymous data isn't necessarily anonymous. Once you cross reference data, you could, for example, build a statistical model and accurately predict the identities* of users. It's been done before and I imagine advertisers do this too (predicting someone's identity* by their web surfing behavior would be an interesting learning task).

* : By identity, I mean some abstract server-side representation of a user (which is not limited to something like a unique number for example).

I think I responded to this already. Given what I know about the advertising industry, I think your fear about this is unjustified. Of course it's easy to do this, but I can tell you that my company and companies we work with won't touch data that's even close to uniquely identifying. Fields we deal with are very broad -- think stuff like "male", and "age 21-28". I'm sure that sometimes this data becomes uniquely identifying, but again -- it's stored in the browser. The information available to a website using HTTP headers sent along by the browser and some fancier stuff from Javascript already makes a browser uniquely idenitfying (link).

I think the more serious problem is when this data is tied to an identifier (like your name, email address, etc.). This is the cardinal taboo in online advertising. It's something that's (successfully, in my experience) avoided at all costs.

You want my support for Google: Be upfront and direct about the costs and practices. Otherwise, I think you're a bunch of hypocrites to your own motto "don't be evil." Surely Larry Page and Schmidt have nothing to hide from us, the unsuspecting user ... right?

If I were Google, I'd much rather lose the business of a few people than disclose everything about how I run my business.

I'd agree that those things would be nice, but I certainly don't expect them.

[iago]

The problem I have is that the cost isn't well known. Nobody thinks about how the advertising agencies work, or how Google and Facebook make their money. And they certainly don't directly disclose their business practices to the users. Instead, users see free software and services thinking that they're free (and they're not). That's pretty deceptive.

Google is pretty open if you ask. Most people just don't care.

And by the way, Google is among many that buries its policies in lengthy legal jargon too. At least it makes users aware of policy changes.

https://www.google.com/privacy - linked on every page, written in plain English.

I can only speculate what kinds of information advertising agencies collect and how they use the data. Whether it can identify me as nslay, my Google ID, a cookie number, an IP address, or a behavior pattern is irrelevant. I don't like the idea that Google, for example, can build (and probably does) an almost complete profile of my web surfing history (since many sites host Google Ads). I personally don't want to be tracked and I am opted-in by default. I have to 1) know that I am being tracked (which is generally kept hidden), 2) Find a way to opt-out (if any).

You don't have to speculate. https://www.google.com/dashboard lists everything Google knows about you. As Sidoh said, it isn't a complete history or anything like that, it's simply the interests that it thinks you have. You can add/remove interests or opt out all together (like I do), after which they'll only give you generic ads for the site you're currently on without any behaviour-based metrics.

And again, as I pointed out: Anonymous data isn't necessarily anonymous. Once you cross reference data, you could, for example, build a statistical model and accurately predict the identities* of users. It's been done before and I imagine advertisers do this too (predicting someone's identity* by their web surfing behavior would be an interesting learning task).

Google defines personally identifiable data as the obvious stuff, plus the not-so-obvious stuff that can later tie back to the user, as you said. They're both considered PII, and they're both protected as carefully as possible and only used as outlined in the plain-english privacy policy.

You want my support for Google: Be upfront and direct about the costs and practices. Otherwise, I think you're a bunch of hypocrites to your own motto "don't be evil." Surely Larry Page and Schmidt have nothing to hide from us, the unsuspecting user ... right?

They are almost entirely transparent, barring, as Sidoh said, the "secret sauce" that would be harmful to release. But they aren't secretive about how they track people, how they choose ads, and the fact that they make money off ads.

[iago]

There's also this:

http://www.google.com/policies/technologies/

And I've done countless training courses that go over those again and again and again.

[iago]

Anyways, the shit that annoys me the most (recently) about Google is the obnoxiousness of their real name policy.

For what it's worth, there isn't really a "real name policy" anymore. This is the message, word-for-word, from the guy that heads that team (and shared with pemission): "We do require that you have a name on your account, but this isn't required to be your name. You can also use initials for your first or last name if you want. It's better if you pick a name that your friends know you by, so that they can find you, but that's entirely up to you."

[while1]

Anyways, the shit that annoys me the most (recently) about Google is the obnoxiousness of their real name policy.

For what it's worth, there isn't really a "real name policy" anymore. This is the message, word-for-word, from the guy that heads that team (and shared with pemission): "We do require that you have a name on your account, but this isn't required to be your name. You can also use initials for your first or last name if you want. It's better if you pick a name that your friends know you by, so that they can find you, but that's entirely up to you."

That's good to hear, but I can't help but feel it's too little, too late.  The damage from their previous enforcement of a real name policy has been already done.  I do remember reading that Vint Cerf had spoken out against the real name policy, so I'm inclined to believe what you say is true.

It's amusing to hear that internally there are complaints about the "Yes" and "Ask me later" YouTube thing.  I'm glad to see that Google employees feel free enough to openly voice their criticisms of absurd decisions like this.  However, the fact that they haven't gotten rid of this yet, makes me wary that Google still is pushing a real name policy to some degree.

[iago]

That's good to hear, but I can't help but feel it's too little, too late.  The damage from their previous enforcement of a real name policy has been already done.  I do remember reading that Vint Cerf had spoken out against the real name policy, so I'm inclined to believe what you say is true.

It's amusing to hear that internally there are complaints about the "Yes" and "Ask me later" YouTube thing.  I'm glad to see that Google employees feel free enough to openly voice their criticisms of absurd decisions like this.  However, the fact that they haven't gotten rid of this yet, makes me wary that Google still is pushing a real name policy to some degree.

Employees are *extremely* empowered to stand up and disagree. It's a really fantastic culture like that. There is a lot of reasoning behind the 'real names' stuff, which I can't really go into. Whether it was a good or bad decision is a matter of perspective, but the Internet at large certainly hated it, and they've responded, albeit slowly.