�ݺ�ߣ

Editor's Notes

• #2: Introduction Goal: Provide an overview of internet marketing and discuss some of the privacy concerns
• #6: First, let’s talk about three basic types of display advertising. “ Display” refers to the type of online advertising that uses banners and images usually seen in the margins of a page. The first type of display ads: Contextual advertising. Contextual ads are banners loaded to a page based on the content of the page. The assumption is that if you are interested in this content, then you may be interested in these contents. The way it typically works is that an advertiser buys keywords from the websites selling the space. Unfortunately, as you can see from these examples, that doesn’t always work out so well. To avoid this, advertisers will often also specify contexts in which their ads should NOT be displayed.
• #7: Targeted ads are ones where the advertisers show ads to you based on data collected on you and stored in your cookies. (More on this data in a bit.) Here, we’re looking at two types of targeting ads. In the top example, Amazon is recommending products to me based on products I’ve looked at in the past. In the second, State Farm is showing me insurance probably based the fact that I have (fairly recently) been looking at new car. In both cases, it was probably my behavior (although it could just be my data) that put me in the segment to be targeted for these ads. This is also called behavioral targeting.
• #8: Another form of behavioral display is called retargeting. Retargeting is done when you visit a site and look at a particular product or type of product and then it shows up in banner ads elsewhere. (The above example is not actual retargeting. I just pulled these screenshots to illustrate how it might look.)
• #9: This image appears in pretty much every presentation you will ever see about online marketing. It’s a diagram by a company called Luma, so it’s called the “Lumascape” of Ad Tech and it shows a lot of the different companies involved in online marketing.
• #10: Let’s simplify the online advertising landscape and do a quick review of how this ecosystem evolved. 1) An advertiser wants to sell you stuff, so they buy ad space from publishers. A publisher sells ad space, which is also called “inventory.” And originally, advertisers would go directly to publishers to buy ad space. 2) With the proliferation of websites, it became too difficult for advertisers to talk to all the publishers in direct buys and lots of publishers had a hard time making money selling space, so they organized into Ad Networks. So, the advertisers could just buy inventory from a collection of publishers all at once. 3) Then, of course, advertisers wanted to be able to reach more people and still not have to talk to a bunch of different publishers and networks. And so Ad Exchanges were invented. These are just like stock exchanges in that publishers are notifying the exchange that a visitor is on their site and advertisers bid on how much they want to pay to have their ad displayed to that visitor. This happens in MILLISECONDS. Very fast. So, lots of times advertisers just have set rules for how much they will pay, or they have computers to execute complex buying strategies for them. So, how do advertisers decide how much they’ll pay to show an ad to a particular visitor? It’s based on the data stored in your cookies.
• #11: Where does data come from? 1) You! The information you volunteer when you register for sites is often used to target ads to you. 2) Your actions. We talked about how your online behavior is used to make inferences about your interests. 3) Offline sources. There are public sources of information, but also as you interact with businesses, they may bring that information online. OPTIONAL DISCUSSION: How does offline data onboarding work?
• #13: Open discussion here. But everyone should be aware, that when you go online, your computer isn’t talking directly to just one other computer. Your computer is talking to manymanymany other computers. OPTIONAL DISCUSSION: How do tracking pixels work? How do cookies get set in the first place?
• #14: Why do you think people are concerned about privacy online at all? 1) They don’t know who knows what about them based on what they’re doing because they don’t know how online tracking works and who is doing it. 2) They don’t want to be embarrassed. [Trey’s opinion: this is the primary motivation among the general public to be concerned about privacy] 3) Some worry that it will be used against them. For example, some are worried about discriminatory practices in targeting ads. Or that they won’t get the best price offered to them: http://lifehacker.com/5973689/how-web-sites-vary-prices-based-on-your-information-and-what-you-can-do-about-it
• #15: In February of 2012, the White House issued recommendations to the online marketing industry. Transparency: consumers should be aware that their information is being used for marketing and what information is being used. Respect for Context: Data should be used appropriately and shared on a very limited basis. Don’t be a stalker! Security: Online data should be protected appropriately. Access & Accuracy: Consumers should be able to see what data is collected and correct it if it is wrong. Focused Collection: Only as much data as necessary should be collected. Accountability: Consumers should have some recourse to pursue when abuse is spotted. While these goals are nice, there are some inherent challenges: - The online marketing industry tries to protect user “anonymity” online. This makes it hard to know EXACTLY who each individual is online. [OPTIONAL DISCUSSION: Third-party vs First-party cookies] - Some data points reflect proprietary/competitive practices - Not all consumers are satisfied opting out or removing their data - Malicious entities will not follow the rules. But the industry does have practices to try to reach for these goals. [OPTIONAL DISCUSSION] Current legislative activity and landscape around data collection, practices, and privacy
• #16: In spite of supposed amazing ability to identify individuals online, compose incredibly detailed profiles, and target them with ads (This is actually not true, per discussion on previous slide.) Why do we still see ads for crazy things? 1) Segmentation algorithms necessarily include an element of probability and inference. Just because it’s wrong for you doesn’t mean it’s not exactly right for a lot of other people. 2) Remnant inventory is left over, unsold ad space. When a publisher isn’t sure what to show you -- perhaps you deleted your cookies or no advertisers bid on you -- then they may choose an ad at random to show. 3) Sometimes marketers make mistakes.
• #17: Here’s where I think there is a legitimate concern regarding one’s privacy online.
• #19: What can you do to protect your privacy online? 1) Delete your cookies 2) You can also change your browser settings to not allow third-party cookies... or any cookies at all. But be warned: this will affect your browsing experience. And if ads targeted to you annoy you, how will you feel about being shown untargeted ads?
• #20: Some companies, like Google and Blue Kai shown here, actually show you the data associated with the cookie they have on you. You can correct this data or delete it completely. These companies usually also offer the ability to opt-out, so they will not target you with ads any more. (That doesn’t mean other people won’t, though!)
• #21: YourAdChoices.com is a program established by various industry groups to educate consumers about online advertising and does also offer a means to opt-out of being tracked by member companies. Pro: you can opt out of lots of companies all at once. Cons: It adds a do not target cookie to your machine, if you delete your cookies, you’ll have to opt out again. Also, not everyone is a member of these industry groups.
• #22: I never delete my cookies. Instead, I am comfortable just being aware of what’s going on and here are two browser plugins that will show you when pixels are firing and for whom they are firing. They usually also allow you to block pixels and cookies wholesale or individually.