How What you Tweet, Tag, or Blog May Determine What Ads You See

In my last post, I wrote about a patent application describing how Yahoo might come out with a widget that could be used with blogs, to recommend old posts on those blogs based upon your lifestreaming activities.

It appears that Yahoo may have even grander and more financially motivated intentions behind collecting information about how you blog, tweet, tag images, and leave other footprints on the Web about your life and interests.

Imagine Yahoo crawling the Web and grabbing information from APIs and feeds published by other sites that provide information about the movies you rent, the reviews that you publish, the pictures that you tag, and the sites that you bookmark. Along with your tweets, your status updates, and your other activities on the Web, this information could be used to build a profile of your actions online.

That profile might then be used to determine which banner ads, job postings, and other advertisements that you may be shown.

It could also possibly be used by some sites to personalize the content that you may see.

That’s the topic of the following Yahoo patent application, which could take lifestreaming data to personalize how you see the Web:

Personalized Advertising Using Lifestreaming Data
Invented by Saurabh Sahni and Pankaj Kothari
US Patent Application 20100023399
Published January 28, 2010
Filed: July 22, 2008

Abstract

This patent discloses a method to increase the relevance of advertisements displayed on the Internet. An ad server may receive a request for an advertisement from a web server. The ad server may compare metadata to online advertisements within an ad database. The metadata may include data about the user obtained from at least two websites through a lifestreaming process.

The comparison may seek out a best match between the advertisements and the metadata and serve the resulting advertisement to the web server.

The patent application provides details of how this system might work as well as some examples that put it into perspective.

Here’s the first of them:

An online DVD rental service shows that Bob likes to rent movies on weekends from an online DVD rental service. Consider a scenario where system has retrieved application programming interfaces (APIs) from a popular online Digital Video Disc (DVD) rental service and those APIs demonstrate that user Bob has tendency to rent movies on weekends.

In addition, system previously retrieved a message posted by Bob on a free social networking and micro-blogging service, where the message as “Watching movie in Mxim Theater, 5800 Zoo Drive, Kansas City, Mo.” In more recent retrievals, system has obtained metadata that indicates Bob has been listening to songs from the movie Screech on a United Kingdom-based Internet radio and music community website. Bob also had identified a trailer of the movie Screech on a video sharing website as being a favorite trailer.

Today, Friday and the first day of the weekend, Bob has requested a webpage (FIG. 1) that details new movie releases and includes a position for an advertisement. With all above personal data about Bob’s activity maintained as metadata, system may use metadata to retrieve an advertisement from ad database that may be closest to “Buy tickets for movie Screech in Mxim Theater.” The ad then may be served into webpage.

The advertising system might draw all of Bob’s activities that it sees together to decide upon the best advertisement to show him from their database of available ads, upon the premise that he would be more likely to click upon that ad.

Another example describes how the actions of “friends” on social networking sites might also play a role in the advertisements that we might see.

The patent filing primarily focuses upon describing how advertisements might be custom delivered to viewers based upon this activity data. There’s only one brief sentence that tells us that other non-advertisement based content could also use this kind of information, and it doesn’t layout much in the way of details:

The lifestreaming metadata also may enable websites to personalize content according to the user activities.

Conclusion

There have been a number of patent filings and whitepapers from the major search engines that describe how search history and browsing history might be used to personalize advertisements. This patent filing from Yahoo describes a way of actively crawling the Web and gathering data from other sources to collect information about individuals and determine what ads to show them.

In What Personalization Means to Search, I wrote about how Search Engines might turn to “analyzing footprints people leave on the Web” to provide personalization to searchers. I also asked what that might imply in terms of privacy.

I’ve been tweeting about shoveling snow this morning. Will I start seeing ads for snow blowers?

What ads might you see?

Share

32 thoughts on “How What you Tweet, Tag, or Blog May Determine What Ads You See”

  1. Hahaha. I had to laugh at the title of the post, because all I could think of is, “What I choose to look at will determine what ads I see”. But obviously many people do look at ads (why they do, I will never know) and for them, this is a service. So, good for advertisers, good for ad readers, great for people who are really into shoveling snow and need to replace an overused snow shovel.

  2. At the bottom of many Google SERPs people are also beginning to see the “Results from people in your social circle for ‘keyword here'”. Google is really driving home the point that personalized search results and advertisements are becoming the wave of the future. It almost reminds me of “Minority Report” where the advertisements are all personalized based on the person.

  3. We’re almost there now, with iphones, and ipads, location-based advertising will be the norm. I don’t think consumers will mind much if it results in lower prices and more selection. The future i see is a move twoards electronic price badges on store shelves that rise and fall depending on time of day, and supply and demand.

  4. This would mean more sales thus more profits. I think it works both ways for customers and sellers alike. For customers, they will at least find ads that are of more importance to what they really need and for sellers because they would be able to target customers that really need their products.

  5. I think this is deeply flawed at best. The idea that my past searches or browsing habits will remain the same in the future. And to display ads based on what I had searched some weeks or months back is simply stupid.
    I agree that it would be great if search engines can “think and analyze” more. But I don’t see that happening any time soon. Also what if I want to browse anonymously? What if I don’t want the internet companies to use my “digital footmarks”?

  6. I think the key to personalised search is understanding the value of what the person has searched for. Recency doesn’t necessarily mean relevancy. Frequency is also a factor.

    I’d be interested to know more details on how the profile information will be aggregated. Using life streaming seems a bit vague to me at the moment, I’m not sure that universal attempts to consolidate social networking presences are sophisticated enough at the moment.

  7. Hi Bob,

    I’m not sure how well consumers or the FTC would react to fluctuating prices on storeroom shelves. Somehow I don’t think that would go over well.

  8. Hi Matt,

    Minority Report was eye-opening on a number of levels. The personalized ads were one of the most frigtening parts of the movie for me.

    I was exploring Google’s social search results, and the people Google had selected to be in my social circle earlier this morning. I’m not sure that I would want my searches influenced by a lot of their choices.

  9. Hi Andrew,

    It is interesting to see how Google and Yahoo and Microsoft are trying to identify ways to target appropriate ads for people who might see them. I still remember a number of years ago when a sponsored ad on Altavista told me that they had Dalai Lama’s on discount when showing me search results for “Dalai Lama.” After seeing that, my thought was that completely random ads might work just as well.

    I suspect there might be many people who would start feeling a little paranoid and creeped out by seeing ads that might match too closely with their activities on the Web.

  10. Hi Jerin,

    I’m not convinced that my past searching and browsing history, and activities on the Web are going to be a good match for the advertisements that I want to see, or often even what I want to see in search results.

    The patent application does point out that they would use more “recent” activities, and not display ads that might be influenced by things that you wrote or tagged or reviewed weeks and months ago. But, often my searches are based upon situational facts that may have little or nothing to do with what I searched for yesterday, or what restaurant I reviewed a week ago, or what images I published or tagged last night.

    Personally, I’ve been starting to use a wider range of browsers, to avoid the impact of customization based upon cookies. It doesn’t provide a complete ability to browse and search anonymously, but it offers some hope of me being able to not have my searches taken over by what the search engines are assuming that I want to see.

  11. Hi FletchtheMonkey,

    The patent application does go into more detail on how profile information would be aggregated, but like most patent filings, it doesn’t provide a full roadmap on how such a system might be set up in intricate detail.

    Given that the related wordpress plugin was developed and deployed in 2008, it’s likely that Yahoo has looked into setting up an advertising system like this more deeply over the past couple of years. If they are using something like this, I’d love to see a whitepaper or some other research that they might have done, that might tell us about how they might address some concerns that you have.

    It’s possible that there may be some additional Yahoo patent applications involving the kind of aggregation of lifestreaming information involved that haven’t been published yet. I’ll be keeping an eye out for them. Saurabh Sahni’s linkedin page mentions his involvement with 4 patents, and I’m only finding two with his name on them.

  12. Hi Roger,

    This is Yahoo’s patent application rather than Google’s.

    That’s not to say that Google hasn’t been introducing more and more amounts of personalization into the search results that they present. In early December, an Official Google Blog post, Personalized Search for Everyone described how they would personalize search results based upon an “anonymous cookie in your browser.”

    They do tell us in that post:

    You’ll know when we customize results because a “View customizations” link will appear on the top right of the search results page. Clicking the link will let you see how we’ve customized your results and also let you turn off this type of customization.

    So there is a way to turn that kind of personalization off, if you want.

  13. Hi David,

    That snow blower would have come in useful. As a commentor noted in my post about the plugin, Yahoo did release a plugin that made recommendations for blog posts based upon lifestreaming data a few months before that patent was filed in 2008. One of the inventors listed on this patent filing was also listed as one of the developers of the plugin.

    Have they been using the same data to determine some ads that they’ve shown since then as well? I’m not sure.

  14. Not only ads are affected. I keep getting the same search results when researching a topic. I am forced to use different search engines to get relevant results. I will also search from a different IP. Someday Google will self destruct.

  15. Some people would probably get all paranoid about their privacy over this. Me, I don’t really mind. I think targeted advertising is better for me as a consumer – I am special, aren’t I? ;)

    Looking at my morning tweets, I’d be getting some weird ads, for sure. It’s been pretty diverse, to the point of “defilers are the little scorpions that make buildings bleed. I’m talking about devourers, the ones that spit purple goo.”

  16. While I do believe that targeted advertising can be a good thing for both consumers and advertisers, won’t it be kind of limiting who sees the ad?

    And what if I start searching randomly from being bored? Does it weigh searches differently to pick the correct ad?

    I do agree with Jerin in the fact that this is flawed.

  17. How will this affect real-time search results that we currently see in PPC ads? Will search history, or personalized search, also show up with regular PPC ads? I guess I’m confused because PPC ads show results for businesses related to my current search, yet Yahoo’s new patent will reveal ads that target my past search history. How will the two be married?

  18. Hi Web Squad,

    I would imagine that many ads might show based upon a specific query used, or the content of the page the ad is found upon, instead of just being shown to specific targeted audiences.

    This advertising process isn’t based upon the pages you select in search results, or your page browsing history, but rather your lifestreaming activities. There are a number of other patents from the search engines involving personalized ads involving your searching and browsing history, though.

  19. Hi Cory,

    This doesn’t involve your past search history, but it might be impacted by your recent tweet, or what you might have taqged an image on Flickr with, or the tags used on a blog post you wrote.

    If you have that information set to go to MyBlogLog, than Yahoo could possibly know about it pretty quickly. Pretty much real-time.

  20. Hi Kim,

    Skeered me, too.

    That’s one of the things that bothered me the most about this patent filing. You never know what someone might tweet, tag, or blog, and it may just not be something you want to see, or something you want to see an ad related to.

  21. I really hate the idea of anything ‘friends’ do online being a part of what I’m served in any way. Just think of that going any further? Sheesh, skeer me.

  22. I think use of the Internet will go more and more this way – it is only natural, if a little scary, that the search engines will try to find out as much as they can about us all so that the advertising they show us is as targeted as possible. The more targeted, the better the conversion and ultimately the more money they will make. I agree with an earlier commenter that I don’t tend to pay much attention to ads that get shown to me, but clearly others do as it is a huge industry.

  23. Ah, its all about ads these days, isnt it…Personally I dont mind too much as long as it doesnt become overwhelming. Ive noticed for the past year or so that the ads I see are more and more tailored to my personal interests and needs. At first it was like “How did they know?” but then it dawned on me “Oh yeah, recall just a few pages ago”. Things are becoming increasingly open and public, Im afraid its nothing that we can control. If this pertains to both sides of the aisle I think its a good development.

  24. Hi MartinTS,

    I agree with you, and I think we will continue to see more and more targeting of ads, but I find myself wondering, “how much is too much?”

  25. Hi Dennis,

    Honestly, I’d rather see ads that are relevant for what I’m searching for, or looking at, at the moment rather than a guess as to what my interests might be.

  26. This really makes you think how much information google really has on each of us, I do like seeing targeted ads more then irrelevant ones, but I can see where this could get dangerous. Facebook had something like this a few years ago and it went way to far, I think they got sued. Good post Thanks.

  27. Hi Cape Cod Website Design,

    Google does have people talking about privacy these days, with the release of Google Buzz, and the fact that it’s tied so closely to people’s Gmail accounts, and automatically adds contacts based upon people that you emailed and chatted with, which can become public if you’re not careful.

    Google does potentially collect a lot more information than that from us, from our searching and browsing behaviors on the Web, our use of Google services, and possible, like Yahoo’s patent filing here describes, the footprints that we leave on the Web in many places.

  28. I just can’t get too upset by Big Brother(s) Google. If someone wants to find me, I’m right here. I’m listing in the phone book, I’m on Facebook and my website domain is openly registered. I have nothing to hide and nothing to fear. Knock on wood. ;)

  29. Hi M.J.,

    The patent I described in my post is from Yahoo rather than Google, but regardless of that, there are a lot of sites that are gathering information about people in ways that wasn’t done before. People share things about themselves that they might not have a few years ago, and those things are out there for potential employers, and acquaintances as well as people who might take advantage of information that people share about themselves.

    I’m not sure if the Web is transforming the way that we think about privacy, but I think it’s changing.

  30. Thanks bill, i actually haven’t enough knowledge over seo. But i’m still learning, learning from bill himself. Really enjoyed your article

Comments are closed.