Google and Behavioral Targeting

Google, so far, has refrained from behavioral targeting. But I think it is about time and it is the logical next action for Google to start offering Behavioral Targeting to its advertisers and publishers. Yahoo and MSN are both testing their Behavioral Targeting capabilities; their Behavioral targeting depends on the user data generated on their own sites. Tacoda and Revenue Science built the concept of network first and then are recruiting publishers to participate in the networks. Google is taking all together a new route. It is busy putting its footprint all over the web. These footprints will help Google build the largest and the best Behavioral targeting Ad network. Google already has publishers (Google Search, Content Network, You Tube) and advertisers (Adwords and PPC). Visitor behavior will come from various Google applications which are everywhere on the web. Google knows more about the user on the web that any other company knows. Google is every where (almost) on the web.
Let’s start from Google Search. Via user’s search keywords and key phrases Google knows what the user searched for, how many times she searched, which sites she visited, how many times and what time of the day she searched. Google might not know the visitors name but knows the visitor via anonymous cookie.
When a visitor arrives at any site from Google search chances are Google will be present there in form of Adsense, Adwords, Google Checkout or Google Analytics.
Even if a visitor by passes the Google search and uses some other way (yahoo search, live search, bookmark, by directly typing in the URL or any other way) chances are she will visit a site which has Google in one or more of above mentioned form.
Google Adsense/Adwords – A visitor who clicks on an Adsense Ad reveals a lot of about her preferences. Just like search Google knows, Google knows which sites (products, offers) the visitor is interested in. How many times the visitors clicks on the ads and what types of ads she clicks on.
Gmail – Google know what emails a user gets, it knows the content of the email, just look at all the ads that show up when you are reading your email. Even if Google does not have the users physical address it knows how to reach her.
Google Checkout – Google knows what a user buys, where she buys from, how often she buys and voila by using Google checkout she just gave Google her name, address etc.
Google Analytics – This is the one of the best tool (as far as behavioral targeting is concerned) Google has put on the web. Not only will it tell Google which sites the user visits, it will also tell Google where she visits them from, what pages she looks at , how long she stays on which site, what she buys, what keeps her engaged and what does not and list goes on.
You Tube, Blogger, New alerts and several other Google products provide will further enhance the data set Google has.
Google Analytics, Google Search and Adsense is where the majority of the data and the power of the network will come from. Aggregated data of all the applications will provide such a rich set of that that within 2 – 3 clicks Google will know weather user is a good prospect for a particular offer, product, service etc. or not.
I think it is a matter of time when Google start connecting the dots and announce it’s entry into Behavioral Targeting. They might call it something else but at the core it will be leveraging the visitors’ behavior all across the web to better target ad on its network.

Follow the Search

Search, as you already know, is attracting a lot of attention from Marketers. People who use search engines to navigate to Web sites far exceed those who type the URL directly into the browser address bar or use bookmarks according to a survey by Piper Jaffrey Investment Research.

Marketers are spending more and more money on Search Engine Optimization (SEO) activities to boost their organic search listings. However, nobody really knows how search engines index various sites and pages in their organic listings. So many times I have seen that a keyword search will show a site’s page (mostly home page) even though that page is very generic and has little to do with the keyword that was searched. The page might have had the content related to that keyword sometime in past or that keyword is still there but there is a lot of other content too, the content that is not really relevant to the keyword that drove the visitors to that page.

Note: Even in Search Engine Marketing efforts, there are a lot of marketers who spend a lot of money buying keywords and then send the visitors who click on their paid listings to a very generic page, most of the time to the home page. Not only are they wasting money on these clicks they are losing an opportunity to convert those visitors into customers. I will cover this in a future article.

A visitor, who types in that keyword and lands on the site, gets confused because he/she does not find what he/she was looking for. Visitors are very impatient, they do not have time to go through all the content on the page to search (yet again) for what they were looking for. As a result visitors immediately bail out causing a very high bounce rate and a lost opportunity for the website owner.

So, as an owner of the site, what do you do?

Simple answer is “Follow the Search” in 5 simple steps. The basic idea of “Follow the Search” is to provide user with a relevant content that will drive them further into your site and hence drive up your conversion. Don’t assume that visitors will find their way because they won’t. Give them an immediate reason to stick around, show them they have arrived at the right site.

Here are simple 5 steps of “Follow the Search”:

1. Capture the search word – As soon as the user lands on your site, capture the keyword user searched on a search engine to get to your site. This can be accomplished by writing a simple code on your page to look for the referring url that drove the visitor to your site. Note: If you have the money then you can also use tool such as Offermatica to do the same.If the referring url contains one of the search engines (you can create your own list of search engines that you want to track but for simplicity I would suggest looking for major search engines and the top search engines driving traffic to your site) then extract the keyword from the referring url. Google, MSN and ASK have the keyword in query string called “q”, while yahoo has the keyword in query string called “p”. Example of the Google referring url is

   http://www.google.com/search?hl=en&q=anil+batra&btnG=Google+Search

2. Build a list of links to the content (or products) that that matches the keyword user searched on the search engine. You can use your on-site search technology or human defined list of content (or products) relevant for your top 50 or 100 (depends on the resources you have) keywords.
3. Use a highly visible area of your home page (or any other page) to display the list that was built in step 2. You can use left side, right side or main content area of the page but make sure it is in the highly visible area of the page, I recommend conduction A/B test to figure out the best location. Use a catchy label such as “Top 10 Resources for [XYZ keyword]” or something similar. Visitors love the top 10 or Top X lists and by tying it with the keyword they searched on the search engine you will make their life so easy that they will not abandon you.
4. When a user arrives from a search engine, capture the keyword (step 1), use the list of content related to that keyword (step 2) and display it to the visitor (step 3).
5. Save the search keyword in session or cookie – Save the search keyword in session or cookie so that you can display the results every time user comes back to the entry page (the page where search engine sent him/her). By saving in cookie you can even show the same results in subsequent visits.

Don’t forget to configure your Web Analytics tool to measure your success – Some of the things you might want to track are

• How many visitors or visits clicked on items in these lists?
• Which links are getting the most clicks?
• What is the conversion rate (whatever your end actions are) by visits (or visitors) who click on these links?
• Change in end action conversion rates
• Change in bounce rates

Here is an example:

Here is a screenshot of http://www.portlandindian.com. This is the page visitors gets when they arrive on PortlandIndian.com. Note, home page above has very little to do with “Roommates”. It has top navigation link and maybe some listings mixed in with classifieds.

When visitors searche on Google for “Portland Indian Roommates”, they get Home Page of PortlandIndian.com as the 1st listing. As I mentioned above this page has very little to do with “Roommates” search keyword that user searched.

However, when the visitors arrives on the site via this keyword, the site follows the above mentioned steps and presents the visitors with the Home page with a section called “Rentals and Roommates” right in the middle of the page.

Questions? Comments?

Presidential Candidates and Behavioral Targeting

I read an interesting article by Michael D Jensen titled “What do the Presidential Candidates use for Analytics?”

One that caught my eye was John McCain, who is using Revenue Science. I don’t think John McCain is using revenue science for web analytics but I think he is using it to participate in Revenue Sciences’ Behavioral Targeting network. Interestingly enough I saw Revenue Science tag on only the home page, other pages did not have any code. Maybe he is still “Undecided” and is testing it out.

Participating in Behavioral Targeting is an interesting concept for Presidential Candidates. In my opinion they can influence a lot of votes by participating in Behavioral Targeting.

Here are some of the ways how the presidential candidates can use Behavioral Targeting (provided they put the tracking code on every single page).

1. Retargeting – If a visitor lands on candidate’s site and then wander over to some other site in Behavioral network, the candidate can advertise to this visitor when they are anywhere on the network. By visiting candidate’s site, the visitor has just shown interest in candiate and so he/she can make sure that this visitor never looses the sight of the candidate(think opposite of out of sight out of mind).
2. Undecided – John McCain has a section called “Undecided”. Any visitor who goes to that section and view 2 or more pages is definitely undecided. Target them with a message (ad) that makes them decide in your favor, this can be used for both on-site and network targeting.
3. Segment Visitors – Segment users based on what content they read or interact with on your site. Using this behavioral understand where they stand in their decision process and then targeting them, on the network or even on-site, with relevant message.
4. Contributions – Say somebody starts a contribution process but never finishes it. Target them with a message that drives them to contribute to your campaign. This can also be done both on-site (after visitor abandon’s the contribution process but still remains on the site) and on-network, follow the user as he/she moves around the network.
5. Use visitor’s off Site Behavior to understand what really makes them tick – If the behavioral targeting network is big enough and have wide variety of sites. Understand which sites visitors visit before they arrive to your site. When they come to your site, show them a message that will align with their off-site (on-network) behavior. Say I care about education and visit sites or blogs (participating in the behavioral network) on education; when I arrive on your site, show me your stance on education. This will help me make my decision.
6. IP Based Targeting – Based on the geo location candidate can
   1. Do more media buy and targeting in the geo locations where they don’t get a lot of traffic from.
   2. Segment user visitor based further and see what the visitors from their top geo locations doing on their site. This will allow you to fine tune the on-site messages.

I am sure more and more presidential will participate in targeting advertising such as Behavioral Targeting.

What do you think? Do you have any more ideas on how they can use Behavioral Targeting? Send those to me.