1. Search Retargeting for Display Ads — lets advertisers target display advertising based on a user’s search activity. So a user that searches on a term like “sandals” could be served a display ad for footwear elsewhere on Yahoo’s network.
2. Enhanced Retargeting for Display Ads — allows advertisers to deliver dynamically generated display ads across the Yahoo network based on user activity on an advertiser’s site. Going beyond standard site retargeting, the new technology would allow an advertiser to target users who visit an airline website to check offers for flights from SFO-JFK, and serve them a personalized offer for that specific flight when they visit a page within the Yahoo Network.
3. Enhanced Targeting for Search Ads — adds capabilities for Sponsored Search and Content Match ads, including ad scheduling and demographic targeting within search. New features are designed to extend the advertiser’s control over where and when an ad is shown at both the campaign and ad group level, including what time of day and day of the week an advertiser would like campaigns to run (ad scheduling), and what age and gender they’d like to reach (demographic). Advertisers will be able to vary their bids for different segments in order to increase their ability to reach the desired audience.

Much of this technology was developed through Yahoo’s BlueLithium division, which was acquired by Yahoo in September 2007.

My question is, why it took so long to get these capabilities available? Google and Yahoo should have been providing contextual and behavioral targeting to marketing for years … especially based on search keyword history and click-through history … Anyways, Kudos to Yahoo to pioneer this space in the search engine battle. Unfortunately for Yahoo, Google maintains control on the search engine space, and growing … almost 82% market share, up 4 points from 77% a year ago.

So what does this mean for Yahoo?! Well, as much as they would like to see is as increase in ad spend, this improvement attracts advertisers and marketers, not end-users, and therefore Yahoo will see a shift in the spend patterns, but not an increase. Yahoo needs apply the same innovation and thought leadership to provide better tools for the end-users and capture back some of the market share they’re quickly loosing.

In the words of Sir Francis Drake (and repeated countless times in pop culture since) “knowledge is power”.

As a person that is painfully enthusiastic about the marketing world, I find it very easy to get caught up in some of the more exciting and innovative work that we’re able to produce across so many mediums. It’s very easy to forget the importance of intelligently addressing your message to make sure the right people receive potentially focused and targeted messages.

All of this messaging can only be accomplished through accurate in depth details on people, businesses, places, and things. Building in depth catalogs on places and things is somewhat easy. Building intelligence on businesses is more challenging as they tend to be private by nature and trying to build in depth details of consumers has become easier and more complex at the same time.

The whole idea for this post started yesterday while chatting with a client. We had their site hooked up to Google Analytics and after giving them a tour they exclaimed, “I can’t believe this is free!”.  I explained that Google, in exchange for giving away analytics software, had received at least two valuable things in return. First, they had a foot in the door to sell you Adwords. Second, they were now collecting deep data on the details of people that visited your site, how the found your site, and how they used your site. Since Google Analytics maintains a dominance in market share of web analytics providers (in terms of the number of sites monitored) they are now collecting shocking amounts of data about the who, when, what, where and why of how people use the Internet.

At the bottom of this post you’ll notice the “ShareIt” component that has become so common all over the web. This common web component now openly posted on hundreds of millions of web pages not only tracks website traffic data, but also stores information on the content that people are most commonly sharing. This brilliant innovation starts to get into a part of analytics that is frequently overlooked: the how and why aspect of people’s online behavior.
I don’t think I need to point out that the social networks are starting to build mind blowing repositories of personal information about people. Not only are they deep into our personal details and individual tastes, but they are also developing an understanding of the friends we keep and how social groups are formed. They might not be selling this information to third parties but they are, without question, using this data to intelligently address the ads within their own network. I don’t think any social network VC firm anticipated their future users would ignore a good portion of their ads (thus devaluing their ad costs to some of the lowest in the ad game), but theoretically they have the ability to bring the most target (and valuable) ads to the table. I foresee the social networks starting to operate in the “gray” area of their privacy policies. They might start to partner with other media outlets using their data to power more intelligent ad serving but continue to keep the data secure on their own systems in accordance with their privacy policies.

Take a company like Google that owns Google Analytics, owns a social network like Orkut, and could easily develop or buy a company such as ShareIt. Now imagine the amount of data that they’re analyzing about the way we live and work.

Ultimately, all of this rich data will power the next generation of the internet, the semantic web, as it evolves over the next ten years. If the current web is the glue that weaves our modern economy together, then dominating the semantic web will be the deciding factor in establishing the most powerful companies of the next decade.

I repeat “knowledge is power”.

Akamai has access to anonymous traffic from all over the world on all type of sites, and has access to track user paths and determine behavior. For example, they can know that a user has been looking at specific cars across multiple sites and suggest a targeted ad on a totally different publisher. The best of all, is that Akamai has access to this data without requiring any integration from publishers, no pixel images, no scripts, just raw data from their content networks.

In addition to their own data, Akamai buys Acerno, who’s specialty is its analytics software which uses anonymous consumer data to find out who’s buying what, who will most probably be interested in a certain product and – most importantly – why.

Combining user traffic, clickthroughs, views, and e-commerce analytical data, Akamai can build a model to serve extremely targeted ads and provide such service to publishers, ad networks and advertisers, and charge a share of the increased revenue or return that the service helps deliver.

Akamai made a point of stating how concerned the company is with protecting consumer privacy, stating that “all of Akamai’s Advertising Decision Solutions, as well as Acerno’s services, rely solely on anonymous and non-personally identifiable end user information”

Behavioral Targeting Advertising
Behavioral targeting is a technique used by online publishers and advertisers to increase the effectiveness of their campaigns. It uses information collected on an individual’s web-browsing behavior, such as the pages they have visited or the searches they have made, to select which advertisements to display to that individual. ISPs route all customers’ traffic, thus they are able to monitor web-browsing habits in a very detailed way allowing them to gain information about their customers’ interests, which can be used by companies specializing in targeted advertising. At least 100,000 US customers are tracked this way, and as many of 10% of US customers have been tracked in this way at one point. Technology providers include NebuAd, Front Porch, and Phorm.

Deep Packet Inspection Background
Deep Packet Inspection (“DPI”) is a computer network packet filtering technique that involves the inspection of the contents of packets as they are transmitted across the network. DPI is sometimes referred to as “complete packet inspection.” Owing to the volume of traffic on most networks, DPI is usually automated and performed by software based on criteria set by the network operator. Deep Packet Inspection can be used to determine the contents of all unencrypted data transferred over a network. Since most Internet traffic is unencrypted, DPI enables Internet Service Providers (“ISPs”) to intercept virtually all of their customers’ Internet activity, including web surfing data, email, and peer-to-peer downloads. After inspecting the contents of users’ packers, ISPs can use DPI to perform activities based on filter criteria. Deep Packet Inspection has been used in attempts to: build profiles of consumers for marketing purposes; intercept communications at the request of law enforcement (both with and without warrants); enforce copyright laws; prioritize the transmission of some packets over others; and identify computer viruses and spam — as described by the electronic privacy information center.

Deep packet inspection — the secret harvesting of granular details about individual internet activity so companies can make better guesses about what to sell you — has been facing a slow death in the U.S. NebuAd, the leader on these shores, has shed employees and lost its CEO amid public backlash and intense congressional scrutiny that led one legislator to question if what they were doing was even legal — greatly quoted by wired.

Critics have likened deep-packet inspection to the phone company tapping a call. The technology allows a window into potentially see all of a consumer’s online activity, from Web surfing and search terms to any unencrypted web communication –says the Washington Post. I’d like to emphasize the term unencrypted, as no deep packet inspection can monitor any activity performed over a secure layer channel, therefore as long as you ensure that all your personal information is always being transmitted over SSL, no middle tear company seating at the ISP can store it.

NebuAd — US’s largest DPI advertiser
NebuAd is an American online advertising company, and it is one of several companies developing behavioral targeting advertising systems, seeking deals with ISPs to enable them to analyze customer’s web surfing habits in order to provide them with more relevant, micro-targeted advertising.
NebuAd’s solution has three main parts: Hardware hosted within the ISP that is capable of inserting content into pages, an off-site server complex to analyze and categorize the contents of users’ Internet communications, and relationships with advertising networks willing to present NebuAd’s targeted advertising.

NebuAd’s privacy policy says that they will “specifically not store or use any information relating to confidential medical information, racial or ethnic origins, religious beliefs, or sexuality which are tied to personally identifiable information”. NebuAd’s former CEO Bob Dykes argues that at the top of that privacy policy list is requiring ISPs to use an opt-in regime when deploying NebuAd’s DPI technology, a notion Dykes said would dilute the effectiveness of the program. “No one, not even the government, can determine the identity of our users”

Due to fallout following public and Congressional concern, their largest customers have all pulled out and as of September 1, 2008, and no ISP is known to be using NebuAd as of today.

Phorm — UK’s largest DPI advertiser
British firm Phorm provides a similar service as NebuAd, but has so far managed to steer through the death-inducing scrutiny and negative press. While Nebuad’s ISP partners have raced for the exits, the British advertising technology company has announced instead that “significant and accelerated” progress has been made toward launching a trial of its online advertising platform with Internet service provider BT Group and has plans to work with ISPs that provide service to 70 percent of the country [says wired], though [free press] announced that a recent ISPreview survey found 57 percent of readers would leave their ISP if they learned it had partnered with Phorm.

Google Chrome — potential root level DPI advertiser
I recently blogged about Google Chrome’s information tapping and their horrible EULA. Days after release, Google announced it would change how long it stores the IP addresses of its users without anonymizing the information, from 18 down to 9 months. This doesn’t change the fact that they reserve the right to monitor, store, and use all information when using Chrome. They were already doing it before with the Toolbar, but this is different; the toolbar only send searches, Chrome has access to every form post, no matter if it’s encrypted through SSL or not.

While Google listens to criticism and modifies their EULA immediately, they still try to justify it as read below:
“While we’re glad that this will bring some additional improvement in privacy, we’re also concerned about the potential loss of security, quality, and innovation that may result from having less data. As the period prior to anonymization gets shorter, the added privacy benefits are less significant and the utility lost from the data grows. So, it’s difficult to find the perfect equilibrium between privacy on the one hand, and other factors, such as innovation and security, on the other.”

The Good
Behavioral targeting enriches the internet on several fronts: Firstly, website owners are offered an improved click-through rate (CTR), which could increase profits or reduce the amount of page-space dedicated to advertising. Owners of previously thought ad-unfriendly websites are offered a chance to make money not on the subject matter of their website but on the interests of their visitors. Secondly, Advertisers are offered better targeted adverts, hence reducing the scattergun approach –publish as many ads as possible in the hope of catching a client– and users are offered more relevant adverts. For last, ISPs get paid for allowing companies like NebuAd access to their network on a per-user per-active profile basis.

The Evil
Although DPI advertisers privacy policy clearly defines opt in/out regulations, it has not been properly executed and users are being monitored and profiled without their approval or knowledge. The technology now only allows to monitor packets, but also to modify them. What does that mean? That DPI advertisers could overwrite the base advertisement affecting the original advertisers, it could insert advertisement, affecting the content producers, and they are doing all this at the ‘last mile’, which means that the content producers and advertisers would never know that someone is stealing their profits. The worst part is that although DPI advertisers claim not to be able to identify an individual and keep all behavioral analysis anonymous, they do track it back to a household, and the ability of removing anonymity is just a click away. The real scary part is that there is a third party company that knows everything about you, which whom you never signed a contract, and has the freedom to do and sell this information to whoever they wish. 

Good or Evil?
DPI can be good and evil at the same time; it has some many advantages to improve experience, but at the same time it is so powerful and scary if it falls in the wrong hands.

Internet publishing networks and other entities have ways of analyze browsing behavior and click-streams of every connected workstation. Every time you open a site, click through another site, search for a product or service, click on a banner or paid search keyword, … every time you do any of those things, someone is watching, logging, and analyzing. You might think it is invasion to privacy; I used to think so too. In reality, they don’t know your identity and they store all behaviors anonymously. Is it bad, awful, a crime? Not really, all they’re doing is improving your browsing experience. The fact of the matter is that if an advertiser buys an ad space, you are going to see it no matter what, but wouldn’t you like of it at least to be relevant? I would. There is no escape, so we better just accept it. To take it a step further, these entities even know to differentiate multiple people sharing a computer in a household by behavioral trends. They will know the difference of dad, mom, little sister, and little brother by the time of day and group of sites they usually visit.

Let’s take the same principle outside, to the real world. Imagine a billboard or posters with tiny cameras that analyze the people walking through. They know who looks, how much time they spend on it, and by facial recognition they even know the age, sex, and ethnicity of the viewers. Imagine if they could record all these data and use statistical information to display targeted ads to demographics by location and time.

Let’s take this last idea a step further. Imagine if these billboards and posters had the ability to store the frequency of your mobile device and capture the fact that you might be interested — this frequency is called RFID and all mobile devices emit it. Imagine if all billboards are connected into an intelligent grid, know if you’re approaching them by following your signal around, and show you direct targeted ads.

Everything I mentioned is either being done or coming. It is rather scary I might say, but none of it is captured against an identity; at least for now, though the scary part is that it could be with minimal modifications.