Privacy and Personalization: From Clickstream to Targeted Advertising

Written by Alex Iskold / August 10, 2007 1:41 PM / 20 Comments

This week we are talking about personalization and privacy is an important, related topic. To personalize pages companies need to know things about us - what we like, what we've seen, what we read and what we purchased. Online retailers are in the business of capturing our information to give them a competitive advantage over their competitors.

On the other hand with the rise of social web we have begun voluntarily putting our private information online. We blog our corporate strategies, share our family photos, and make lists of our interests. Yet most people are still concerned about privacy. They may not know exactly what it is that makes them so concerned, but they are concerned. In this post I will look at privacy in the context of personalization from user concerns to target advertising and ownership of clickstreams.

Why Are We Concerned About Privacy?

Many times over the past few years I had conversations where people asked: But what about privacy? My answer is always: What exactly are you concerned about? The majority of people just worry about privacy as a word; they can't express what it is that worries them. It is a conservative, mostly uninformed behavior: "I just don't want them to know about me."

This is a rather naive position. First of all, "they" do know about you and you agreed to it when you did not uncheck the privacy box. So a better worry would be to find out who they are and what they are going to do with your information. Obviously, you do not want your information to be used against you. And that is the real reason that people are concerned about privacy.

It's Not You, It's Your Money They Are After!

The good news about the people stalking our online behavior? They don't want to hurt us, they just want our money. The reason retailers want to know our private information is because they want sell us things.

Google is the reigning king of the web because it solved a simple problem - how to effectively advertise to users. As it turned out, the answer was embedding ads into the search results. But the search engine is not the end, it is merely a means to an end. It is the advertising that makes the whole system work.

Google ads, however, are contextual, but they are not personalized (at least not yet). Google delivers advertising based on the content of your search results, which certainly works well, but could work even better if the search results, and the ads, were tailored to the specific end user. This is not possible to do without knowing your personal information.

The Behavioral Targeting is Personalization

In our earlier post this year, The Art, Science and Business of Recommendation Engines we discussed the main ways that personalization engines like Google, Amazon and Last.fm work:

Personalized recommendation - recommend things based on the individual's past behavior
Social recommendation - recommend things based on the past behavior of similar users
Item recommendation - recommend things based on the thing itself
A combination of the three approaches above

Effectively Google is an item or context-based recommendation engine and Last.fm is the social recommendation engine. Amazon has a combination of all approaches, but most notably, Amazon has recommendations based on your past experiences, such as what you've previously purchased, rated, or looked at.

The problem for Amazon's competitors is that practically no other retailer has the kind of information that Amazon has to do this kind of personalization. To solve the problem, we have seen a range of Behavioral Targeting solutions. All of these solutions utilize clickstream information to show you a targeted ad.

For example, companies like Tacoda create user profiles by serving and tracking cookies from hundreds of web sites. Using sophisticated algorithms they assign users to different categories by interests, but they never identify you specifically. Based on the collected information they are able to show you and other people like you relevant advertising.

Tacoda enables relevant ads and from that point of view it is the same as Amazon - both try to maximize clicks. The difference is really in how the private information is being used. In Tacoda's case, the information is anonymous, it does not know that you are really you. But Amazon does.

Who Owns The Information?

This brings us to the crux of the issue - who owns our attention or behavioral information? Earlier this year, I attended the Open Data workshop in New York City, where influencers from the industry spent a day debating this issues. The split was about even, 50% of the attendees thought that tracking people via cookies is perfectly fine, while the rest thought it was wrong.

Intuitively, the information belongs to the consumer, but when we look into the details, things become less clear. We explicitly choose to use Amazon, to click and to buy things there. Everything we do is a two way street, since Amazon provides a service and we transact with it, it seems that they should have a right to the data as well.

In case of the Tacoda, things are also not simple. Tacoda's technology is installed on the sites that you are visiting. In the fine print on those sites there are disclaimers. You click and transact with these sites just like you do with Amazon. Why should this be any different? It seems that the sites have the right to leverage your information.

Conclusion

Privacy is a very controversial topic. Everybody has an opinion and everyone is concerned about how their information is being collected, stored, and used. It is particularly odd to hear privacy concerns and then login into Facebook and see people putting everything about themselves in their profiles. But such is the day and age we live in.

To effectively settle these privacy issues and the dispute over ownership of information we need laws. Without specifics, all we get is what it says in the Fifth Amendment of the US Constitution: nor shall private property be taken for public use, without just compensation. But does that include clickstream? That is truly a billion dollar question.

Comments

Subscribe to comments for this post OR Subscribe to comments for all ReadWriteWeb posts

Alex -

One of the better explanations regarding data privacy that I've seen - thank you. Nothing is free and the cost of free content is advertising. In order to serve up relevant ads, I pay by relinquishing information about my behavior. The issue is that this agenda is hidden.

In most cases it is VERY difficult to "turn off" tracking and almost impossible to erase history once you find out what data has been captured/tracked. What is missing is not privacy, but control over privacy - who gets to use my data and who doesn't. And how secure is my data?

I believe the Contextual Web is a good thing, but I want to transparency rather than obfuscation around how that data is used and some control over who has access to the data and where it is stored.

As my company plays in the Contextual Web space, I follow this topic very closely and have a pretty strong opinion on the subject. In the end, it is about getting ME, the consumer to spend MY money with a product or service vendor (advertiser). It's about time that all of these companies thought about who their END CUSTOMER is and design a solution around them.

Thank you,

Liz

Posted by: Liz Coker | August 10, 2007 3:20 PM

Hi Liz,

This is true, it is about hidden agenda. This is nothing new, though, it has always been like this in the business world.

The problem is, as you are pointing out, is that it is almost impossible to opt out. Like you can't tell Amazon do not suggest. Their answer is - this is what we do.

As individuals we feel very strongly about this, we do not want to be watched - period. But, companies feel that each visit is a two way street and they have a right to learn about us to maximize their profits.

Its a tough subject for sure ;)

Alex

Posted by: Alex Iskold | August 10, 2007 3:30 PM

Of course I agree that many internet companies just want to make money.

But if companies can get deeply personal data about you, then so can governments. A government has only to tell its own legal system to force a company to turn over whatever a company has.

Obviously not all companies and governments are bad, but the only way to make sure they're not bad is for them to be transparent.

I'm not ignoring the question of whether or not one can actually secure information about oneself.

In fact, it's this combination of lack of transparency in governments and the level at which our personal lives are being mined that has me uneasy.

Posted by: Eric Eldon | August 10, 2007 10:39 PM

I think the APML (Attention Profiling Markup Language) is a good approach to control what's our possible privacy's degree

The first step is to know our profile, the part of us we send around the Web and so try to trace and store our attention data, practically spy ourself.

Posted by: Federico Bo | August 11, 2007 1:25 AM

Eric,

You raise very good point, this is probably should be a subject of another post.

Alex

Posted by: Alex Iskold | August 11, 2007 5:02 AM

Hey People, just wake up. This issue falls in the heart of private property rights. The data that is collected by the search engine vendors about you, is their private intellectual properties and they can do whatever anything about data as long as they are not violating any law, that might land them in court.

If you don't like using search engines then don't use one. I can hear some that would say, Oh, but it is my right to use a search engine. Yes, perhaps a government owned search engine that you have the right to its use, since you pay tax that contributes in funding that state-owned search engine. For the Google search engine or similar ones (Yahoo, Microsoft, etc), it is a private property, which is not state-owned, and all you whingers here have no right to demand what Google , Microsoft, Yahoo, what they can or can't do with their own private properties , ie, collected data about you.

Hands up those here who disagree with this or think that they have a right to someone else property? If you think you have a right to demand Google, not to do anything with the data that they collected about you, then perhaps, you list your address here, so that I can drive around your neighbourhood and take a good look at your house, examine it closely and see if I don't like its look (its architecture or its color). If I don't like your property perhaps for its ugly look, then I will demand you that you should paint it according to a color of my (not yours) choosing.

Now, if you don't like the hypothetical scenario that I have described above, then it is exactly the same as Google feels. Whatever they want to do with their own property such as collected data (their own house) is none of the public's business. Likewise, it is none of my business to come and demand you that you should re-paint your house with a different color since I don't like its look.

Posted by: Falafulu Fisi | August 11, 2007 5:19 AM

Alex as you should know by now - the conversation has moved on a great deal from these topics.

The tools now exist for users to literally convert their lifestream/clickstream into Attention Profiles - and the standard exists for the Attention profile to be ported from service to service.

You can refer to www.Engagd.com and www.APML.org for more info.

The transparency and ownership now rests with the user - users and app developers now need to start demanding it from the web.

Posted by: Chris Saad | August 11, 2007 8:09 AM

Chris,

I do not think you really read the post. It has nothing to do with attention. Attention information is only one small aspect of privacy issues. In this particular case, the focus is on tools that automatically collect information when you click around on the web.

You keep on asking me to look at APML. I have and you know that, as well as you know what I think. I was very clear in my post on Attention Silos. I would like to see these standards evolving around an organization with a bit broader involvement from industry experts and technologists.

Alex

Posted by: Alex Iskold | August 11, 2007 3:41 PM

As you say the focus of the post is about privacy as it relates to collecting information as you click around the web. That information is commonly called Attention Data.

In regards to standards, Standards form where they form. It's not a perfect system. The wonderful value of the web is that anyone can propose and implement something and the community decides what to adopt. Sometimes it comes from great and dedicated standards bodies, and sometimes it just comes from people trying to fill their own needs.

APML has a broad and growing set of support in the form of the APML workgroup/mailing list and various discussion groups.

If you know of an active body who can adopt and broaden the conversation we are all for hearing about it.

In the mean time it's time for a little less conversation and a little more action.

Posted by: Chris Saad | August 11, 2007 4:28 PM

Chris,

It is ironic that you are calling me to act.

I belong to handful of people who are pushing the attention rock up the hill and trying make things work by actively blogging about it, talking to industry leaders, big companies and through my work at AdaptiveBlue. It is a complex issue and is not going to be solved via one incomplete spec.

I like what you are doing, but I would suggest that you should not mix your agenda with public good - this is not the way to get things done.

Alex

Posted by: Alex Iskold | August 11, 2007 4:44 PM

@Alex - Regarding the question of law, while I'm 100% sure the 5th amendment was intended for questions of imminent domain, I'm only 50% sure that there needs to be a legal solution to this problem. In cases like this, I can see google and yahoo taking leadership roles in educating the public to the exact ramifications of their actions online, thus avoiding a need for a legal solution. That said, where google and yahoo have been powerless to educate or prevent misuse of the Web (i.e. Spam) the law has tried to solve the problem. The results are unsuccessful. Again, I'm 50% sure. I still think that online consumers of content deserve to retain their anonymity as a basic legal right and if a law is needed to preserve that right, so be it.

@Chris + Alex - I agree with you both. APML, while new, is laudable in its approach to attention transparency. The spec needs help, yes, but it is a step in the right direction. In fact, I see it and the use/adoption of it as a great start in educating consumers of online content how their "data" is used by the sites they visit. In an ideal world, this personal attention data is transferred on an opt-in basis where consumers of content are driving the adoption up to publishers and service providers (and not vice versa).

Posted by: Kelly Abbott | August 11, 2007 7:35 PM

@Kelly - I agree with you. But I personally think its less about "privacy" and more about "control". Who has control over a piece of information is where I think the conversation should pointed. If users are in control of their own personal data (attention based or otherwise) then they are able to protect it however they deem appropriate.

Since Web 2.0 has effectively handed control from publishers to users, I think that this is a simple step to extend the power of the user. i.e., if publishers abuse my personal data, then I will simply "vote with my feet". I think that the fear of loosing audience, at least in the web 2.0 world where audience is so valuable, will keep publishers and service providers in-line.

Posted by: Ashley Angell | August 11, 2007 8:15 PM

@Ashley - Good points. Control is very important. And empowering. Furthermore, empowerment can enable privacy. Let's complete that tautology: control == privacy? :-)

Posted by: Kelly Abbott | August 11, 2007 8:42 PM

One of the problems is precisely our lack of control. Web 2.0 gave us tools to expose our attentions. But it did not give us tools to preclude people from capturing our information as we browse the retail sites.

Alex

Posted by: Alex Iskold | August 11, 2007 8:44 PM

Alex I think you do a great service by keeping the Attention discussion in the public conciseness by posting it to a mainstream blog like R/WW. Well done. However I think it is just as important to acknowledge the efforts being made and move the conversation forward. Perhaps you could even support them with your own projects (you mentioned AdaptiveBlue for example).

I, for one, would love to know about active open standards and tools that are emerging to capture and remix Attention metadata from all the vendors out there in the ecosystem. Who are the handful of people you speak of pushing the Attention Rock up hill? Let the world know of your valiant efforts!

I have posted more on the Engagd blog.

http://blog.engagd.com/2007/08/controlling-your-own-attention-data.html

Also, I would welcome your opinion on our own efforts with Engagd.

Posted by: Chris Saad | August 12, 2007 1:12 AM

1) Chris and Alex - please stop. You are both extremely intelligent and doing great things with your efforts in your respective domains. I would like to see less bickering and more discussion on the themes. We've only scratched the surface.

2) I am a member of the APML workgroup, and I work at a firm that has 140,000 people and makes $20 billion annually. The enterprise architect's are well aware of APML and the need for attention management, because I introduced them to it and I know they like it. In fact, I joined the workgroup because it's the best effort I have seen to tackle the issue. Having said that though, APML is only part of the solution.

3) If we are going to argue about it, we might as well formally define it. Start with my definition -
http://liako.biz/2007/04/define-privacy-what-does-it-mean-to-you/

Posted by: Elias Bizannes | August 12, 2007 3:33 AM

Chris, what does APML intend to achieve? What sort of profile does it mean to capture? The term 'Profile' has a very broad definition and I am not sure what benefits APML do bring to users.

Posted by: Falafulu Fisi | August 12, 2007 6:27 AM

Falafulu, to quote from the APML.org website:

"APML allows users to export and use their own personal Attention Profile in much the same way that OPML allows them to export their reading lists from Feed Readers.

The idea is to boil down all forms of Attention Data ‚Äì including Browser History, OPML, Attention.XML, Email etc ‚Äì to a portable file format containing a description of ranked user interests."

Posted by: Chris Saad | August 12, 2007 8:39 PM

Great post Alex. Getting people over the naive first reaction of - "I'm scared about losing my privacy" is important for the industry and you have produced the simplest and clearest explanation of how the issue should be thought about that I have seen.

Posted by: Nic Brisbourne | August 13, 2007 2:47 AM

A strange reaction of this author's post about APML.
In the blogs are common to talk with a large vision on the subject's post.

Posted by: Federico Bo | August 13, 2007 11:57 AM

ReadWriteWeb