Friday, August 29, 2008

Search Behavioral Targeting

Microsoft and Google both have plans for using users past search behavior to provide tailored search results, practice commonly known as Behavioral Targeting. While companies like Revenue Science and Tacoda (now part of AOL) use users web surfing behavior on the internet to determine what users are interested in and server Relevant Ads, search giants are using the past search queries and click though activity to determine users intent and show them relevant search results (organic and paid ads).

As these practices become more prevalent no two users will see the same results on search engine results page for same queries, some of this is already happening. In future even the same user might see her results change from one day to another based on what he/she searched and clicked on between those two days.


An example of how search behavioral targeting might work

A user looks for hotels in Bahamas and then search for Scuba Diving lessons the next day, she might see the search results (organic and paid) that combine both the queries to show her Scuba Diving Lessons in Bahamas. Yet another day she decides to search for Jamaica. Guess what? Her organic and paid results might show her Hotels, Airfares (though she never searched for it) for Jamaica and maybe Bahamas or even some other Caribbean destination. She might also get results for local Scuba Diving lessons.
So why will the search engine showing all these results when user never explicitly searched for them?
Search engine is trying to determine user's intent based on what user searched for and then show the customized results based on the intent. In this case the user's intent is clear; she is looking to go somewhere, most likely in Caribbean. Since she is searching from a geo location other than Caribbean, she might need flight and hence airfare might be shown. She has already indicated she needs Hotel and is looking for Scuba diving lessons so her search results are related to those keywords as well. Once the intent is determined search engine can show can show customized results for a particular user. As the user searches more her intent become even clearer and the possibilities of how search results can be customized for this user are endless.



Last July Google started using immediate search queries to show targeted paid search results. Google is now expanding the time frame that it uses to look for search queries. Here is what NYTimes.com reported on Google’s plan:

Nick Fox, a director of product management who looks after ads on Google’s search site, said the company was now testing the use of more search queries in its ad targeting. He did not describe how it was doing that. But Internet experts said that it was most likely using its cookies.
Mr. Fox said that Google’s approach was different from what Yahoo, AOL and others call behavioral targeting. Those companies look at what a user did a few days earlier to show them ads about the same topic today. Google says it believes that search engine advertising is most effective if it relates to what the user has most recently searched for.
“We are trying to understand what the user is trying to do right now,” Mr. Fox said. “In some cases, those queries are ambiguous, so you need a little more context.”
Google’s previous system of looking at a user’s immediate past query was not useful enough, he said. “It is probably not just the previous query that matters,” Mr. Fox said. “You want to know if the user is still doing the same thing. You wouldn’t want to go back a month. You wouldn’t want to go back a day. But you may want to go back two or three queries.”


Microsoft plans to take it one step further and expand the “behavioral targeting” to the organic search results. Here what is what CNET reported:

Specifically, the company believes examining a full sequence of user queries can lead to more useful results. Today, the company only keeps track of the immediately prior search, but often users use search engines to explore subject areas broadly, said Satya Nadella, senior vice president of Microsoft's search, portal and advertising platform group, at the Search Engine Strategies conference.
"I believe this notion of understanding user intent--being able to analyze (search queries) and come up with search patterns and use them to shape the search experience--is one of the most important areas for us," Nadella said.

Recently Behavioral Targeting has been under a lot of fire from privacy advocates and lawmakers. It will be interesting to see how privacy police reacts to search behavioral targeting. Though Behavioral Targeting from Revenue Science is anonymous, Search Behavioral Targeting is based on user provided information (search keyword), which can easily be tied to PII information.

I have always advocated an opt-in model for Behavioral Targeting, a system where users give their consent to be tracked and have an easy option to turn the tracking on or off. I believe that Search behavioral targeting makes it even more desirable to have an opt-in model.

Comments? Questions?

--------------------------------------------------------------------------------
Looking to fill your Web Analytics or Online Marketing position? Try WebAnalytics Job Board

New Position

(Web Sales) Conversion Marketing Manager at Hewlett Packard (American Fork, UT)

---------------------------------------------------------------------------------

Wednesday, August 20, 2008

A/B and Multivariate Testing Landscape

In our most recent whitepaper we investigated the current state of A/B and Multivariate testing, focusing on how the practice is currently used and the obstacles companies face in implementing this type of an online strategy.

Our analysis, taken from surveys given at the eMetrics Marking Optimization Summit, found that 52% of online marketing managers are currently engaged in A/B or Multivariate testing and that an even greater percentage plan to begin testing within the next year.

A high-level overview of the findings:

  • A variety of testing platforms are considered before making a final decision

  • Experiments vary in frequency, page type and page element

  • Practitioners often use more than one platform to meet their needs

  • Pretest hypothesis often fail

  • Lack of Budget was cited as the main reason by those who currently do not do any testing

  • Complexity of Tools and lack of best practices were two top challenges faced by those engaged in Testing

  • A/B and Multivariate testing is “worth it”

  • Behavioral targeting is a small but growing practice



Surprisingly lack of budget was one of the top reasons why companies were not involved in A/B or Multivariate Testing. But if you look at the chart below you will see that companies that do conduct A/B and Multivariate testing find it that it’s worth it. Majority of the respondents who engage in testing said that either it is worth it or can’t live without it. So if you are one of those who think because of lack of budget you can’t do A/B MVT, think again? Your Return will make it worthwhile to do testing. You will be able to move budget from other areas such as Ad spend to A/B testing. Do your ROI calculations. You will be surprised with the positive results you will get with developing a culture of continuous testing.








You can download the whitepaper from ZeroDash1 website.


--------------------------------------------------------------------------------

Looking to fill your Web Analytics or Online Marketing position? Try WebAnalytics Job Board

New Position

Web Data Analyst at Alzheimer's Association (Chicago, IL)

---------------------------------------------------------------------------------

Monday, August 18, 2008

WebTrends Appoints New CEO

Management shakeup at Webtrends continues. The shakeup started last year when Greg Drew left Webtrends. After Greg, Webtrends appointed Bruce Coleman as the interim CEO, till it found Dan Stikler as the new CEO of the company. Today Webtrends announced the departure of Dan and appointment of Alex Yoder as the new CEO.

Here is the press release from WebTrends:

The Board of Directors of WebTrends Inc. today announced the immediate appointment of Alex Yoder as the company's new president and chief executive officer following the resignation of Dan Stickel. Yoder was also appointed to the company's Board of Directors.

In naming Yoder as CEO, the board turned to a WebTrends insider with seven years of experience at the company, and who most recently led WebTrends to its most successful quarter in history as its vice president of sales. He also has been one of the key executives leading the company's long-term strategic planning, and has over 20 years of global business experience.

Commenting on the management change, board member Neil Garfinkel of Francisco Partners said: "Alex has a proven track record and has demonstrated clear leadership at the company over the years, helping to form our strategic vision and drive customer success. He’s the right person to lead WebTrends’ business forward.”

Yoder joined WebTrends in 2000 and has served in various management capacities in the sales organization. From 2007-2008, Yoder was the vice president of sales at Touch Clarity, a behavioral targeting company, before leaving to rejoin WebTrends as vice president of sales.

"I've spent seven of the last eight years focused on expanding the WebTrends brand and delivering business value to our customers,” said Yoder. “I'm excited to partner with other WebTrends leaders such as Nick Sharp, vice president and general manager WebTrends EMEA, to achieve our next phase of growth and become the clear worldwide leader in marketing optimization solutions. We will achieve this with an unwavering commitment to product innovation, by delivering valuable and compelling solutions to our customers and by building strong relationships with our partners."

WebTrends recently announced its highest revenue quarter in the company’s history, and tomorrow will unveil its new search engine marketing optimization solution—WebTrends Ad Director™.


My best wishes are with Alex and team and I hope that now Webtrends will be able to find the stability and growth that is long overdue.

Sunday, August 10, 2008

Opt-out from Google and Yahoo Ad Network

Online tracking and advertising based on users’ online behavior have got a lot of heat and scrutiny from privacy advocates and lawmakers.

As a result of this scrutiny Yahoo and Google/Doubleclick are now both providing an easy way for user to opt-out of ad targeting on their receptive networks. Yahoo and Doubleclick were part of Network Advertising Initiative (NAI) and allowed users to opt-out from their network via NAI’s opt-out tool.

Google

To find out more about Google network cookie and how to opt-out visit http://www.google.com/privacy_ads.html

Since a lot of you are working in the field of web analytics, you must be wondering how google opt-out will affect the data tracked in web analytics. The simple answer is that this option will not affect Google Analytics. A first party cookie from the site using Google Analytics is used for tracking user behavior on the site using Google Analytics, this cookie is separate from Google ad network cookie. According to Google
“A different cookie is used for each website, and visitors are not tracked across multiple sites…. To disable this type of cookie, some browsers will indicate when a cookie is being sent and allow you to decline cookies on a case-by-case basis. “


Yahoo

Yahoo has been offering that opt-out option for the ads the company runs on it outside partner sites in its network. Yahoo will now extend opt-out option to ads displayed on its own sites. You can read more about this option.
Below is excerpt from Yahoo Press release

Anne Toth, head of privacy and VP for policy, said, "Yahoo! understands the trust of our users is our greatest asset, so we strive to create the most trusted, compelling online experience."
"Yahoo! strongly believes that consumers want choice when customizing their online experience and they have also demonstrated a strong preference for advertising that is more personally relevant to them," continued Toth. "However, we understand that there are some users who prefer not to receive customized advertising and this opt-out will offer them even greater choice."
This new opt-out capability is expected to be available for consumers by the end of August. Users will be able to access the opt-out in the Yahoo! privacy center, which is linked on the home page and nearly every page on the Yahoo! network. Users will also be able to access the opt-out through a link in the public service advertising campaign Yahoo! has been running with online ads across its network to educate users about customized advertising.

Yahoo and Google have taken the steps in right direction but they are not perfect. As I wrote before, both these models are dependent on an opt-out cookie. If you opt-out of these networks and later delete your cookies you will again be automatically opted-in. I have advocated an opt-in model for Behavioral Targeting. This model will remove this dependence on cookie for opt-out. I do realize publishers and ad-networks concern that opt-in model will limit the reach. It is possible that the opt-in model might limit the reach initially but in long run if the value proposition is strong for users then user will opt-in.

What do you think?

--------------------------------------------------------------------------------
Looking to fill your Web Analytics or Online Marketing position? Try WebAnalytics Job Board
New Position
Web Data Analyst at Alzheimer's Association (Chicago, IL)
---------------------------------------------------------------------------------

Sunday, August 03, 2008

Career in Web Analytics

Web Analytics is one of the hottest fields these days. A lot of people are planning to switch careers and many students are looking to start their career in Web Analytics but don’t know where to start. This article is for aspiring web analysts to help them make a transition into web analytics field.

Skills required to be a Web Analyst

First and foremost you need desire and passion to be a web analyst. Desire and passion will get you where you want to go. I believe (and this is my opinion only), if you have the desire and passion then you can acquire other skills. Not everybody will agree with me but again, that’s my view.

The other most important skill that you need is analytical skill. If you are a person who always looks at the problem from a different angle than most of the other people, you have what it takes. If you can put different pieces of the puzzle together to form a complete picture you have the skills to be a web analyst. If you can critically look at things, you have the skills.

Other skills and education that will come in handy are:
1. Business
2. Marketing
3. Statistics
4. Technical

You don’t need a college degree but a lot of employers look for it and I look at it when hiring a candidate. Business, Marketing, Accounting, Statistics and Technical degrees will be very helpful in getting you the job but I have seen Web Analysts with very diverse educational background.
I highly recommend reading the series of interviews I conducted with several web analysts. This will give you an idea of what they think are the skills required to be a web analyst. They also have advice for people who aspire to a career in this field.

Checkout what other web analysts think about the skills required to be a web analyst.

Learning about web analytics

There are several resources available to learn about Web Analytics. There are several blogs on Web Analytics where you can get all levels of information on this subject.

First and foremost you should join WebAnalytics group on Yahoo. This forum is a great source of information. You will find all levels of web analysts in this forum. This is a free for all forum, even if you want to stay on the sideline and just read messages, you can learn a lot. If you have any question on this subject, feel free to ask at this forum.

Buy a copy of Web Analytics Demystified by Eric Peterson and a copy of Web Analytics: An Hour a Day by Avinash Kaushi. These are the two best resources for all levels of web analytics, especially for those who are just starting. Here is a list of books recommended by other web analysts

Online Course by University of British Columbia

If you are prefer to learn in formal way then I recommend the course offered by University of British Columbia. You can learn more about this course at http://www.tech.ubc.ca/metrics/curric.html. (Full Disclaimer: I am one of the associate instructors for this course). This course is offered in partnership with Web Analytics Association (WAA).

Online Reading

Reading blogs, articles and whitepapers is another way to expand your knowledge. There are several blogs out there which touch on various topics on Web Analytics and other related topics. I would recommend you familiarize yourself with the Web Analytics field (see above) before reading these blogs. Both of these blogs have a list of lots of other blogs on Web Analytics. I won’t give you a list of top 10, top 20 blogs because it all depends on what topic you are trying to learn more about. The more you read the better you will understand this field. If you do not have time to go to each and every blog every day you can setup blog alerts on Google for “Web Analytics”


Gaining Practical Experience

This is one of the biggest challenges facing people who want to start a career in Web Analytics. Employers don’t want to hire you if you don’t have experience and you can’t get experience if you don’t have job. (See Getting a Job for more information)

Google Analytics has made it really easy for anybody to get a web analytics tool. This tool is completely free with all the documentation to help you get rolling. If you have a website, deploy this tool and play with it. This will help you understand how web analytics tools and reports work. To gain further experience, tap into your network, I am sure somebody (a friend of a friend of a friend…) will allow you (especially if you are willing to do it for free) to provide reporting and analysis on their site (real site).

Getting a Job

Bloggers and analysts write about the shortage of Web Analysts all the time. There are plenty of companies who are looking for web analysts but are not able to fill the open positions. But when you look at open positions they all need at least 2+ year of experience in omniture, webtrend etc. So what do you do when all you have is some Google Analytics experience on your blog and a certificate from UBC?

As I said before, you don’t need any experience to find a job in this field, all you need is analytical mindset and desire to learn and grow. There are companies who will hire you, train you and provide the support to help you grow. Even though the job requirements asks for 2+ numbers of years in web analytics or web analytics tools, the fact is that there is indeed a shortage of experienced web analytics people. So the companies might be willing to settle for someone without any experience as long as the candidate can demonstrate analytical thinking, desire and passion, and of-course a certificate from UBC will help. Chances of you getting a job without any real experience are pretty good. However, keep in mind that these companies might not be able to relocate you, so you will be limited by the opportunities in the companies in your geographic area.

My advice is that if you see such a job opening, go ahead and apply. If the company has a senior analyst who is willing to teach then you will get a call.

During the interview process make sure that there will be someone who is willing to provide you on the job training. That will make your life easy. You will probably learn faster than those who don’t have that kind of support.

Also, make sure to understand what your role will be. I am amazed to see that some of the job postings want everything on the earth: Omniture, .Net, SQL, Marketing Analysis, Business Objects and the list goes on. If you do apply and get a call from these companies, make sure you clearly understand what you are getting into and that’s what you want to do.

Go ahead and email me (batraonline at gmail) your resume and I will see what I can do for you.

If you have any questions please feel free to email me or a leave a message with your contact information.

Note: This post is an updated version of the blog post that I wrote in 2006.


--------------------------------------------------------------------------------
Looking to fill your Web Analytics or Online Marketing position? Try WebAnalytics Job Board
New Positions


---------------------------------------------------------------------------------

Wednesday, July 23, 2008

Mobile Analytics: Omniture and Bango

Earlier this year I predicted a few developments in Web Analytics and Behavioral Targeting Space. One of them was the rise of Mobile Analytics.
Here is what I predicted:


Mobile Analytics – Mobile sites and mobile advertisings gain momentum. Mobile Analytics will be become a requirement to understand how the sites and advertising are performing. Currently you can either use log files, WAP gateway logs if you are one of the mobile carriers (this is what I did back in 2001 -2002) or services from Amethon. This year we will see more players in mobile Analytics. The reporting will become better than what has been offered to date.

Two recent press releases are proving that I was right in my prediction, one is from Bango and another from Omniture.

According to destinationCRM, a research has found that between 20 percent and 30 percent of overall Internet traffic is generated by mobile devices. In this article Matthew Langie, Omniture's senior director of product marketing said that the need for a mobile-analytics solution was clear: Clients were requesting this capability, and research alone has shown that consumers are readily using mobile phones, with subscriptions at 3.9 billion in 2008 and predicted to reach 5.6 billion by 2013. (Individuals, it's worth noting, may have multiple subscriptions.) A recent study by market research firm IDC predicted that 164.7 million smartphones will be shipped worldwide this year -- a figure expected to reach 363.2 million in 2012. Moreover, while 2007 saw 27 million phones with Wi-Fi capabilities, the service will reach 400 million by 2012.

According to a research by Bango, the most important feature requested by mobile site owners was “The daily/weekly/monthly number of unique visitors count”

However, capturing an accurate number of unique visitors is not easy in Mobile sites as not all of the mobile phones accept cookies and users share the same IP address (network address) and this IP also changes as users roam. The new offerings from both these providers claim to provide an accurate count of unique visitors. It remains to be seen how accurate they are. It is interesting that both of these providers announced their offering on the same day.

Bango

Bango unveiled new mobile site analysis capability to help website owners measure the value of mobile sites by accurately counting unique visitors browsing their site. This new capability in Bango Analytics, v3.0, complements existing campaign analysis functionality .
Bango uses a privacy protected Bango User ID to count and track unique visitors. The Bang User ID is compiled through sophisticated WAP gateway profiling, data from browser analysis, session information and network interactions. This unique user ID enables Bango to distinguish between new and repeat users and therefore precisely quantify the number of unique visitors.
Source : MarketWatch

Omniture

Omniture announced availability of Mobile Analytics in Site Catalyst version 14.1. According to an article on destinationCRM Omniture's effort to address the most common problems marketers face when tackling the mobile arena includes the following four features:


  • In a partnership with mobile Internet advocate group dotMobi and its DeviceAtlas database, SiteCatalyst communicates with an extensive library of device profiles to accurately identify the mobile device accessing a Web page. This determines what video, audio, and text can be displayed, thereby giving marketers insight into what customers are using most frequently. Moreover, carrier identification gives marketers the ability to partner with carriers to set up other campaigns.

  • Through the use of visitor identification technology such as cookies, subscriber identification, or header information, SiteCatalyst can identify a new visitor or a repeat visitor to deliver an experience akin to the traditional Internet.

  • The solution aims to capture the right data from a variety of devices to garner insight about the page. Omniture's server places on each Web page a very small file that occupies a single pixel, allowing marketers to track the activity of the mobile device: where the user is clicking, what content is being downloaded, what video is being watched, etc.

  • A geolocation component that recognizes where the site is being accessed from, thereby enabling Web publishers to deliver advertisements applicable to that location.


As I predicted earlier this year, Mobile Analytics will gain more momentum in the coming months and I will be writing more on this subject.

Thank you Sarah Dawson for helping me with this post.
--------------------------------------------------------------------------------
Looking to fill your Web Analytics or Online Marketing position?
Try WebAnalytics360 Job Board
New Position Mktg Web Analytics Manager at NetApp (Sunnyvale, California)
---------------------------------------------------------------------------------

Goal Attribution to Organic Keywords - Google Analytics Tips and Tricks

In April I wrote a blog post to show you why some of your keywords show 0 Visits and 0 page views in your Google Analytics Report. In this post I am going to show the attribution of goal to the search engine keywords, when a user searches multiple keywords on the search engines to visit your site (all within the same session) and converts from one of the keywords.

Note: The past post and this one are both based on the Organics keywords searches and clicks.

I conducted two following two experiments
  1. Converted on the Last Keyword

    1. Searched “1 page no register seattleindian” on Google, arrived on http://www.seattleindian.com/ , viewed one page and existed the site by typing Google.com in the browser address bar

    2. Searched “2 pages no register seattleindian” on Google, arrived on http://www.seattleindian.com, viewed 2 pages and then exited the site by typing Google.com in the browser address bar

    3. Searched “4 pages register seattleindian” on Google, arrived on http://www.seattleindian.com , registered on the site (converted, Goal 1), viewed total of 4 pages and then exited the site by closing the browser

    4. All of the above was done within 30 minutes and using the same browser session


    Visit and Page View Attribution

    As you can see my visit resulted in 3 keywords, total of 7 page views and 1 visit (visit time out is 30 mins and all of it was done in 30 minutes). As I showed you in the last post, 1 page view is shown and attributed to 1st keyword. The other keywords do not get visits or page views attribution (Figure 1). The total pages are accounted and counted in the keyword report even though 6 pages were not attributed to any particular keyword (Figure 2)


    Figure 1 (click on the image to enlarge it)



    Figure 2 (click on the image to enlarge it)


    Goal Attribution

    In this scenario, the Goal is attributed to overall search engine keywords but not to any particular keyword.


    Figure 3 (click on the image to enlarge it)


  2. Converted on the First Keyword

    1. Searched “SeattleIndian 4 pages register test 3” on Google, arrived on http://www.seattleindian.com , viewed 4 pages, registered on the site (converted, Goal 1) and then exited the site by typing in Google.com in the browser address bar

    2. Searched “SeattleIndian 3 pages no register test 3” on Google, arrived on http://www.seattleindian.com, viewed 3 pages and then exited the site by typing Google.com in the browser address bar

    3. Searched “SeattleIndian 1 pages no register test 3” on Google, arrived on http://www.seattleindian.com , viewed 1 page and then existed the site by closing the browser

    4. All of the above was done within 30 minutes and using the same browser session


    Visit and Page View Attribution

    In this case I converted (Goal 1) when I arrived via the first keyword. When I look at the Site Usage of keywords, the first keywords is credited with 1 visit and 4 pages, the other two keywords did not get any credit of the visit or the pages that were viewed as a result of click on those keywords. So the 3 pages are not attributed to any keyword. This is what I showed in my last post.


    Figure 4 (click on the image to enlarge it)



    Figure 5 (click on the image to enlarge it)


    Goal Attribution

    In this scenario when the conversion happens from the first keyword, the goal is properly attributed to that keyword.


Conclusion

When a user searches multiple keywords to arrive to the site,
  1. The visit is attributed to the first keyword only

  2. The page views directly related to the first keyword are attributed to that keyword and other keywords show 0 visit and 0 page views

  3. Total page views from all the keywords are counted in the overall keyword report

  4. If the conversion happens as a result of the first keyword then it is attributed to that keyword

  5. If the conversion happens as a result of any of the keyword other than the first one, then the conversion is not attributed to any of the keywords

  6. The conversion from any keyword is counted in the overall keyword report


What’s next? I will be testing how attribution works when a user clicks both Organic (SEO) and Paid PPC (search result) within the same visit.

--------------------------------------------------------------------------------
Looking to fill your Web Analytics or Online Marketing position?
Try WebAnalytics360 Job Board
New Position Mktg Web Analytics Manager at NetApp (Sunnyvale, California)
---------------------------------------------------------------------------------