Archives for July 2007

Ad Targeting on Google Search Engine

I wrote about Google and Behavioral Targeting extensively on this blog. Recently I found some examples of Targeting on Google Search engine and would like to share those with you.
Last Sunday I was looking at Keywords Report on Google Analytics for http://www.usaindian.net/. I saw a key phrase “bay area desi car buy” which drove few users to the site (Desi” is a term used by east Indians (India) to refer to each other). I got curious and went to Google to see where this site ranked on Google for this key phrase. As I was researching the keywords on Google, I noticed 3 types of paid ad targeting on Google Search Engine
1. Geo Targeting – Which we know that user allows it advertisers to do
2. In Session search targeting (I made up this term)
3. Combination of Geo and In-Session search targeting

I am going to show some screen shots that will show you these targeting in action

  1. Geo Targeting – I searched for “bay area car buy” (notice, I removed the word “desi” from the original query as it was not relevant to what I am going to show”.)
    1. Got paid ad related to my query “bay area car buy”
    2. I live in Seattle area and got an ad related to my geo-location as well.

  2. In-session Targeting – After seeing the above results I got curious about what other kind of targeting Google is doing so I changed my query to search for “Chicago Car Buy” and found
    1. A paid ads related my search query “Chicago Car Buy”.
    2. I also got paid ads related to my previous query “bay area car buy”.


    I confirmed that this was indeed happening by doing yet another search “Atlanta car buy”

  3. Combination of Geo and In-Session targeting – After my previous search on “Atlanta car buy” I went back to search again for “bay area car buy” and found
    1. An ad related to my query “bay area Car buy”
    2. An Ad related to my previous query “Atlanta Car buy”
    3. An ad related to my geo location “Seattle”

    If you like this post, you might want to subscribe to my blog feed.Click here to subscribe to Web Analysis, Behavioral Targeting and Online Advertising

Consolidation in Behavioral Targeting Space: AOL buys Tacoda

In my 2007 Predictions I predicted
“Only few main behavioral network players will be left and some of the existing ones with poor networks will either go out of business or be sold.”

Well, since then we have seen consolidation in this market earlier this year, Google-Doubleclick, MSN-Aquantive, Yahoo-Rightmedia now the latest is AOL-Tacoda.
According to NYPost’s article Aol has an Eye on Your Business, AOL is moving into the behavorial-targeting ad market with a deal to purchase Behavioral Targeting Firm Tacoda.

“Tacoda, which will operate as a wholly owned AOL subsidiary, is one of several online ad firms that use “behavioral targeting” techniques to track Web surfers’ habits. Through its so-called “Audience Network,” Tacoda helps marketers tailor ads to individual users for such things as cars or computer equipment. “
“…While not referring to any company in particular, [Andy] Falco did concede to The Post that Tacoda “did have other choices, but they choose to come to AOL.”

I am sure this puts Revenue Science (Tacoda’s biggest rival) in sort of an odd position because AOL was one of their biggest customers and were probably competing with Tocoda to be acquired by AOL.

Next acquisition target: Revenue Science. We will have to wait and see who buys them? Any takers? Valueclick?

If you like this post, you might want to subscribe to my blog feed.Click here to subscribe to Web Analysis, Behavioral Targeting and Online Advertising

Web Analytics Education Opportunities

Web Analytics adoption is gaining momentum but there is a lack of skilled web analysts, according to a report by Aberdeen Group
“Web analytics usage has reached mainstream status with 82% adoption among companies surveyed recently by Aberdeen. However, a vast range of maturity exists regarding analytics process, data analysis and corporate understanding of web metrics. A fundamental impediment precluding many companies from building a successful analytics program is a lack of skilled employees required to manage, distribute and analyze web analytics.”

This is exactly what I predicted earlier this year in my “Predictions for 2007”, I wrote:
“There will be a lot more jobs in this [web analytics] field in 2007. A great year for those who are planning to enter this field or looking to move into better jobs in this field. Most marketing jobs will have web analytics as a requirement. Currently there are 1024 open job on Indeed.com but I expect this number to rise as there will lot more openings than qualified candidates.

Since then I have been covering about Web Analytics Job market and, as I predicted, the open web analytics positions are up from those listed on January 1st. On July 1st, “Web Analytics” jobs listed on indeed.com were up 54.79% from Jan 1st numbers (All the past articles on Web Analytics Job market are filed under Web Analytics Jobs on this blog). I expect to see even higher number of open positions in coming months.

So how do companies and aspiring web analysts educate themselves if they want to learn about Web Analytics? This question is answered by the two part reports titled “Web Analytics University” (Part 1 and Part II) by Aberdeen Group. In this report they have reviewed all the educational opportunities currently available.

According to these reports, there are four educational avenues currently available

• Vendor Sponsored Programs
• Analytics Consultants, Blogs and Guru Sessions
• Community Forums & Industry Associations
• Academic Programs

I was flattered to see my blog listed in “A sampling of prophetic blogs and bloggers”. The blogs listed in this list are

Lies, Damned Lies… (Ian Thomas)
Occam’s Razor (Avinash Kaushik)
WebAnalytics.be Blog (Aurelie Pols)
Web Analysis, Behavioral Targeting and Advertising (Anil Batra)
VisualRevenue – Web Analytics & Affiliate Marketing blog (Dennis R. Mortensen)
Web Analytics Demystified Weblogs (Eric T. Peterson & Judah Phillips)
Web Metrics Guru (Marshall Sponder).

Master of Science in Analytics

Now you can get a graduate Degree in Analytics, according to this report,
“A new program launching this month by North Carolina State University’s Institute for Advanced Analytics will offer a Master of Science in Analytics (MSA) degree. Program director Dr. Michael Rappa believes that the best education he can provide is a foundation of analytics that will span online and offline channels and enable graduates to contribute to building the business processes that make up a foundation for an analytics culture within
an organization.”

“Although the program prerequisites include general subjects such as calculus and statistics, Rappa found that nearly 60% of prospective students already had advanced degrees, including 1/3 that were already MBAs. Out of the hundred applicants that bid for the inaugural class, 25 were enrolled, which is astronomical for the first year of a new degree program according to Rappa. What’s even more astounding is that major corporations are already seeking to hire the entire graduating class, even before admitting the first student!”.

This is pretty cool, I was not even aware of this program.

Great reports!!! I highly recommend these reports to all those who want to learn about web analytics.

If you like this post, you might want to subscribe to my blog feed.
Click here to subscribe to Web Analysis, Behavioral Targeting and Online Advertising

Last Call to participate in Web Analyst Salary and Skills Survey

I am conducting a simple 5 minute survey on “Web Analyst Salary and Skills”. I have had a good response so far; however I need some more responses. If you have not taken the survey then please take 5 minutes (or less) to answer few simple questions.

Thanks to all of you who took the survey. I would also like thank Manoj Jasra, Marshall Sponder and Stephane Hamel for spreading the word about this survey (If I missed anybody then please email me their/your information).

I will appreciate if you can spread the word about this survey among your colleagues and friends.

I hope to publish the results of this survey by end of this month (or early next month).

Increase in subscribers with longer gap between posts

Ryan Turner has a great post titled “Going Viral is not a Marketing Strategy” on his blog.

In this post, he says

“Anil Batra noted a while back that his number of subscribers increased when he had a longer gap between blog posts. Goodness gracious! What could that mean? Apparently Anil has a theory. Yo Anil, care to share?”

He is referring to the pattern that I had noticed in terms of posting frequency and number of subscribers to the blog. I observed that that when there was a longer gap between my postings the subscribers to my blog increased far more than when I was posting regularly. So why was it happening?

Here is my theory

Visitors see the blog, like the blog, bookmark it (they reason why they don’t subscribe are listed below) and come back to check it. Since I write regularly they come back directly or via bookmark. If I keep writing regularly, visitors make it a habit to come back and check for the new content. But if there is a break in writing, they still want to read what I have to say, they just don’t want to keep coming back only to find that there is no fresh content. So they resort to subscribing to the feed (via feedburner link and a form that is on the right hand side panel).

I think the reasons why some visitors don’t subscribe to the feed blog in the first place are

  1. It is probably much easier to bookmark the blog and come back to check.
  2. The location of the subscribe form and link are probably not highly visible (right panel), even though most blogs have them at similar locations. To overcome this problem I should have something like the following in every post.

“If you like this post, you might want to subscribe to my blog feed.

Click here to subscribe to Web Analysis, Behavioral Targeting and Online Advertising

I will try above link in next few posts and see if that makes any difference.

Now I need your feedback:
Readers of this blog, I would like to hear your views on this theory.
1. If you are subscriber of this blog then when do you subscribe, first time when you visited, after few times when you got interested in what I was writing or when you saw that there was nothing new for few days?
2. If you read this blog regularly and have not subscribed then what have you not subscribed? Is the subscription link hard to find? Any other reason?

Bloggers, have you noticed similar pattern? What is your theory?

I will further explore the traffic patterns, subscriptions and gap in the postings. If I find anything interesting I will post them on this blog.

Freelance Webanalytics

I am pleased to announce “Freelance WebAnalytics”, a blog for web analyst seeking freelance work and companies seeking freelance web analysts to help them with small projects.

Why this new blog?

I get occasional emails from readers of my blog who fall in one of the following three categories

  1. Web Analysts who are seeking some more work to make few extra bucks.
  2. People who have read my article Starting a Career in Web Analytics and want to learn about web analytics. These are the people who are willing to work for free so that they can learn.
  3. Those who are looking for some help and don’t have budget to involve a big consulting firm.

The idea of this blog is to bring those people together.
Hope this blog will help.

So if you are looking for freelance work or looking to hire a freelancer please send me your requirements at batraonline at gmail (dot) com and I will post them on the blog.

Bounce Rate Demystified

Time and again my clients ask me about Bounce Rate. This made me think that there is still confusion about what is bounce rate and exit ratio. The three main questions that have come up are

  1. What is bounce rate?
  2. What is the industry standard for bounce rate?
  3. What causes high or low bounce rates?

I am going to answer these questions in this post.

What is bounce rate?

Bounce rate is the percentage of visitors who enter a site (or a page) and then leave immediately. Think of a ball (visitor) that is thrown (visits) towards a table (site). It hits the table and bounces back without rolling (visiting any other pages).
Generally, “leave immediately” in the above definition means without going to any other page. However it could also be expressed in terms of time spent on site, say users who spend 5 seconds or less on the site irrespective of the number of pages they view.

Bounce rates are calculated both at the individual page level and at the site level. For an individual page, bounce rate is the ratio of visitors who enter the site from that page and leave without going any deeper, to the total number of visitors who enter the site through that page. In other words it is single page visits/ total entries to the site through that page.
At the site level, bounce rate is simply single page visits/total site visits.

Note: If a visitor enters though a page, refreshes it (manually or via auto refresh such as ESPN score page or MSN money page) but never goes beyond the first page the visit is not counted in the bounce rate.

  • Bounce rate is often confused with Exit Ratio. Exit ratio is usually expressed as the percentage of exits from a page to the total number of visits to that page. As a side note: A lot of times exit ratio expressed as % of visits can be misleading. In most cases, page views are actually more appropriate than visits for this ratio. Why page views and not visits? If I view the same page twice during the same visit, and after one of those page views I exit, shouldn’t my exit ratio be 50% rather than 100%? The first view of this page was compelling enough for me to further engage. A 100% exit ratio would indicate a problem that may not be there.
  • Bounce rate is confused with Single Page Visit Ratio: Single page visit ratio is calculated as a percentage of single page visits over total visits to a page.
    1. Here are two examples that will help you clarify

      1. A visitor who enters site at home page and then goes to contact us page and leaves from contact us will be counted in the exit ratio from the contact us page but won’t be counted in the bounce rate of contact us page.
      2. A visitor who enter the site from contact us page and then leaves without going any further counts in all three, exit ratio, single page visit ratio and bounce rate.

      So to Recap:

      Single Page Visit Ratio= Single Page Visits to the page/ Total Visits to the page
      Exit Ratio= Total Exists from the page/Total Visits on the page or Total Exits from the page/Total Page Views of the page (see explanation above)
      Bounce Rate= Single Page Visits to the page/Total Entries to the site through that page.
      Note: All of the above three are generally expressed as percentages.

      What is the industry standard for bounce rate?

      The simple and short answer is that there is no industry standard. I know you don’t want to hear that, but it is true. There is no industry standard. There are some ranges that I will share shortly but we can’t call them industry standards. There are a lot of factors that influence the bounce rate, so you really can’t compare bounce rates of one site (or page) to another. I have listed those in the next section.

      The goal of the site should generally be to reduce the bounce rate to as low as possible. The lower the bounce rate the better job the site is doing to keep users engaged. One exception may be a site that is intended to accomplish all relevant user engagement on it’s landing page. This is more common on say, a campaign landing page intended to sign up users for direct marketing emails.

      Bounce rate is very unique to your site and page. The best way to know if you are doing better or worse is to set your own baseline and compare your performance over time.

      I have seen most bounce rates fall between 18 – 30% on home page and the site overall. Any page with a bounce rate higher than 30% should be looked at closely. I am not saying that you should not analyze the pages below 30% bounce rate. Remember there is always room for incremental improvement.

      There are several factors that determine the actual bounce rate of any page.
      Here are some of the numbers that were listed by Steve Jackson based on his experience with various sites.

      Source: http://tech.groups.yahoo.com/group/webanalytics/message/6116

      Retail sites driving well targeted traffic 20-40% bounce.

      Simple landing pages (with one call to action such as add to cart) I’ve seen
      bounce at a much higher rate, anywhere from 70-90%.

      Content websites with high search visibility (often for irrelevant terms)
      can bounce at 40-60%.

      Portals (MSN, Yahoo groups etc) have much lower bounce rates in our
      experience 10-30%.

      Service sites (self service or FAQ sites) again usually lower 10-30%.

      Lead generation (services for sale) 30-50%.

      Per Steve,
      “I must stress that all the above figures are based purely on our own
      experience after working with clients. I wouldn’t advise you base an
      optimization model around these numbers. We advise that when forming a
      benchmark, that you do it internally. Take the average bounce rate over a
      given period on your current site. You need to have at least 1000 entries
      coming from normal sources to get reasonably actionable data.
      Measure what the average bounce rate is and then work to get that down.”

      What are the factors that affect the bounce rate?

      Below are some of the factors that determine the bounce rates. You can use this as a checklist to diagnose a high bounce rate issue.

      1. Source of your traffic – Each source results in a different bounce rate. When setting your baseline create overall baseline and baselines for each traffic source e.g. display advertising, organic traffic. With one client I found out that the traffic driven by searches (paid and organic) and sources other than campaigns had a much lower bounce rate than traffic that was driven via display ads. Their display ad had 90% bounce rate while other traffic only had 35% bounce rate. Their overall bounce rate was around 55%, way lower than 90% and giving them a misleading picture.
      2. Search engine ranking of the page – A page which ranks higher on irrelevant keyword will get a higher bounce rate. I have seen this to be an issue a lot of times. I wrote an article on how to follow the search and reduce your bounce rate.
      3. Type of Audience – If you are advertising and reaching the wrong audience you will see higher bounce rate. Bounce rate will tell you if you need to better target your ads.
      4. Landing Page Design – Landing page design affects the bounce rate. I suggest A/B testing to improve after you have set your baseline. No matter how low you go there is always an opportunity for improvement unless you somehow achieved 0% bounce rate.
      5. Ad and Landing Page Messages – If the messages on your banner or search ads are not aligned with the messages on the landing page then the chances are you will have one of those 50% + bounce rates. Make sure messages are aligned and give visitors a clear call to action. Many a times I have seen marketers sending users to a generic page instead of an appropriate landing page. This can (and will) result in higher bounce rates. Again A/B or multivariate testing should be used to reduce the bounce rate.
      6. Emails and Newsletters – Subject lines, to and from, links, banners, the layout of email and the landing pages all work in tandem. They can either result in a great user experience and hence lower bounce rate or can result in a disaster. Do testing (More on this later in another post) to reduce bounce rate.
      7. Load time of your page(s) – A longer load time can result in visitor bailing out of the site causing higher bounce rates. Conversely, users can hit the refresh button, thinking there was a problem with the page load. This will incorrectly reduce bounce rate.
      8. Links to external sites – A page that has links to external sites (or sub domains/ pages that are not tracked in the same data warehouse) will show higher bounce rates.
      9. Purpose of the page – Some pages’ purpose is to drive users inside the site while other pages provide the information that user is looking for. A page that provides the end result can show higher bounce rate. One example is the support page on my bank’s web site, I have this page bookmarked. Whenever I need my bank’s phone number, I go to my favorites, pull this page, get the number and leave.
      10. Other factors – Pop-up ads, pop-up survey requests, music, streaming video, all can have an adverse effect on bounce rates if users become annoyed.

      Hope this clarifies the confusion around Bounce Rate. I would like to thanks Brad Gagne, who challenged my thinking on this subject, provided his valuable insight and proof read this article.

      As I was writing this post I came across another article on the same subject by Avinash Kaushik, you can read his article at http://www.mpdailyfix.com/2007/06/bounce_rate_sexiest_web_metric.html.

      Here are the other articles in the Demystified Series

      Question? Comments?

      Sponsored Message: Reduce Bounce Rate and Increase Conversion Optizent UnBounce. Convert any page into a unique landing page on the fly.

      Web Analyst Interview: K. Trinadh Rao

      Continuing my series of interviews with Web Analyst, here is an interview with K. Trinadh Rao. Trinadh lives in Chennai, India.

      What is your current position and the name of the company you work for?
      Currently I am with Franklin Templeton as “Manager – Web Analytics”. Basically I am responsible for defining the process in report metrics that are usable for different stake holders. I manage relationships between eBusiness and global stakeholder teams for web reporting in Franklin Templeton. Also, as an organization stand of point of view, Franklin Templeton is really serious on setting up internal competency and helping the business with meaningful reports and analysis.

      How long have you been working in this field?
      I have been in the field of Online Marketing & Analytics for more than 6 years. I started my career as Web specialist, responsible for intranet & internet sites. Later I moved on to measuring the site usage and other online marketing activities like SEO, SEM, and E-mail marketing and finally into pure play web analytics and strategy. Interesting part of all these online activities is that they are linked with Metrics to show their ROI and how well they integrated with the website/online property.

      Tell me about your work, education prior to making a switch to Web Analytics.
      I started my career with Ausvista in Chennai, later moved to CollabNet. After CollabNet, I worked with Cognizant for a short stint, where I was responsible for setting up Web Analytics competency and sell the service to Global Clients. I had a chance to meet and present the true value of web analysis to Global Clients from various verticals/domains. Worked with different business units to identify and measure key performance indicators (KPIs) and goals. I did my bachelors in Maths and Web Analytics course from UBC.

      Why did you decided to switch to pure play Web Analytics?
      It was after I met Tim Cloonan (Ex-Director of Marketing, CollabNet). He was so keen on measuring website success and other online campaigns. This gave an opportunity for me to consider Web Analytics as a key factor that drives online business forward. Initially I had to work on Log files for Web Analytics. We created a web interface that helps Marketing & Sales folks to see the web reports directly on weekly/monthly basis. Later, I had a chance to work on different tools like ClickTracks, Google Analytics, WebTrends, NetTracker, and Omniture for different assignments.

      How did you find your new job? How long did it take?
      I think it was from Linked In or Monster. Not sure exactly! I have been receiving tons of offers from different companies. Finally I have decided to work with Franklin Templeton which is a Global Investment Management company and I am interested to work on financial domain. It took me 3 weeks and 7 rounds of discussion to get into Franklin Templeton!

      What are you job responsibilities? Describe your typical work day.
      Right now, I am responsible to oversee web report production for all Franklin Templeton eBusiness managed sites as well as specified intranet and service sites. As I mentioned earlier, I develop and manage processes for reporting; report requests, requirements gathering, distribution and communication. I represent the interests of Website Reporting to Franklin Templeton eBusiness team and act as a leader for web analytic knowledge and statistics inside the organization.

      What are the skills that you think are important for a web analyst?
      I believe that an ideal Web Analyst should poses blend of Business and Technical skills. He should be articulate and understand the online business quickly.

      What education or work experience helped you in making this transition?
      I feel my previous experience working with companies that treat online marketing and reporting as critical business component, helped me to continue in and contribute to Web Analytics field. Offcourse, UBC helped me a lot!

      What education is lacking, education or experience that would have helped?
      If anything truly lacking, it is the awareness of Web Analytics. I am trying my best to create this awareness among the people here by conducting WAW.

      What web analytics/online-marketing books have you read and/or own?
      Web Analytics Demystified. I read a lot from blogs and articles from people like Jim Sterne, Eric, Avinash, you (Anil) and Hurol

      What helped you in finding your new job?
      I would say my experience in this field helped me to get a new job!

      What are the major challenges that you are facing in this industry?
      I find lot of companies still not able to figure out what kind of reports they need. I agree there are a lot of data sets available now-a-days, but making meaningful reports from the data and get actionable insights from the reports are the key challenge. Also, decision making is major challenge.

      How do you make sure you are learning and growing in this field?
      Continuous learning and keep monitoring on industry trends and growth. I will make sure to spend 1-2 hrs daily on web analytics group mails, and blogs from Eric, Avinash and you (Anil).

      Do you have blog?
      In process of setting up one, very soon you will see that!

      Any advice to aspiring web analysts?
      Pick up the web analytics basics first! Try to understand the business models of various internet companies and analyze how web metrics can help them in driving their online business forward. Last but not least, be passionate about web analytics!

      Thank you for your time Trinadh.

      If you are in web analytics field and would like to interview for my blog please contact me at batraonline at gmail.com. Please take a 5 mins survey that I am doing to understand web analysts salary and skills.

      Web Analytics Salary and Skills Survey

      I have been writing about hot web analytics job market and the shortage of web analysts in the market. As a result of these posts, I am getting lot of queries asking if I have any data on the Web Analyst salaries.

      I also get a lot of questions regarding the skills required to be a web analyst. I did a post while back on “Starting a career Web Analytics”and posted my views, Stéphane Hamel wrote an article recently on the same subject. But what is missing is you all’s input.

      I am taking this opportunity to get your input. Please take 5 minutes to take this simple survey to help me understand the salary and skills of web analyst. I will provide the results on this blogs later this month.

      Bloggers, I will appreciate if you can help me promote this survey.

      Thanks for all your help.

      Yahoo Gets Smarter with SmartAds

      NYTimes reports that Yahoo is about to announce a product called SmartAds that would allow marketrs to create custom ads on fly based on visitors interests based on their demographic and online behaviors or actions.
      The articles says
      “For example, a person who had recently searched for information about blenders might see an ad from Target that gives the prices for the blenders that are on the shelves in the store closest to that person’s home.
      Last week I wrote a piece on targeting passions. In that post I wrote
      “However the more granular you get with your segments the reach becomes a problem. Yes there will be quality but reach might not be enough to justify the cost and efforts of reaching these targets.” One of the biggest factors in the “cost and effort” was the cost and effort for creating different creative. Seems like Yahoo’s SmartAds will resolve this issue…pretty cool. We will have to wait and see how it works, if it really delivers on it promise than the marketers dream of 1:1 marketing is one step closer.

      According to NYtimes, this is how Yahoo’s new system works: the advertiser (or its agency) would provide Yahoo with the components of its display ads — including the logos, tag lines and images. The retailer would share information from its inventory databases that track the items on the shelves in each of its stores. Next, Yahoo would combine that data with the information it has about its users’ demographics and actions online to create a product-specific advertisement.

      According to the article: “For airlines, SmartAds uses Yahoo’s information about its Web surfers to create display advertisements for each person that feature ticket offers with actual prices listed. In time, Yahoo plans to offer rich media advertisements where users can buy the ticket at that price right within the ad unit, rather than having to click through to another Web site.”

      Web Analysts, here is something to start thinking about – offsite transactions. User will never come to you site (your web analytics tools, omniture or webtrends won’t even see the visitor), any web analytics tracking embedded within Rich Media will possibly be Yahoo’s.

      Dear Publishers, this will allow yahoo to build richer set of data, advertisers make sure you own that information and not used by yahoo to empower your competitors or against you e.g. raise the CMP or CPC for you because you have higher conversion. I don’t think that should be the case but there is always a possibility.

      Side Note: According to Reuters “SmartAds will operate on Yahoo’s network of Web sites as well as in its advertising partnerships with eBay Inc., major U.S. newspaper publishers and cable operator Comcast Corp.”. Not sure if eBay, major newspapers and Comcast are just the publishers of advertising or will actually be pooling in user behaviors. If these publishers also allow yahoo to create better understanding of users based on their behavior on their own site, yahoo will have even wider reach and better targets. Example: A user comes to ebay looks at some cars and then browses over to a newspaper site, without even clicking on a single link on the newspaper site, Yahoo (network owner) will know this is somebody who is interested in cars. By pooling in behaviors eBay will be able to generate extra revenue stream and the newspaper will have a better target and hence higher CPM.