Thursday, December 14, 2006
One of the short coming of Web Analytics tools (and various marketing applications) is their inability to provide a complete view of all their marketing efforts and provide a way to take actions on their visitor’s behavioral data and help optimize user’s experience. To get a 360 view of your visitors you had to do a lot plumbing and even that did not provide you the full view as each system provided different numbers for same data points. This plug-and-play interface promises that marketers will be able to easily integrate between marketing applications.
I expect to see other web analytics tool providers either getting into similar partnerships or making some acquisition. This is the future of Web Analytics. Web Analytics is not complete if you can’t fully understand various factors that drive traffic, understand what visitors do when they arrive on the site, segment these users based on their behaviors, demographics etc and then take action i.e. provide content, products and experience that is relevant for your segments. These partnerships will allow marketers to get closer to 1:1 marketing.
This is a good start, now we will have to wait and see how Omniture delivers on this promise. I would love to hear your views and experiences.
Wednesday, December 13, 2006
Contributing to this series are:
- Eric Peterson
- Marshall Sponder
- Gary Angel
- Avinash Kaushik
- Anil Batra (me)
- Justin Cutroni
- Jason Van Orden
- Robbin Steif
- Akin Arikan
- Manoj Jasra
Yesterday he featured Avinash and I, you can read the full article at http://manojjasra.blogspot.com/2006/12/identifying-solving-client-pains-part.html
The two most important pain points that I have encountered over and over are:
1. Accuracy of the Data - The tools are purchased and set up without understanding the goals of the business. Since the goals of the business/site are not properly understood clients start measuring and reporting on whatever out of the box reports the tool can provide. In most of the cases tool is capturing information that should not be captured and skipping information that should be captured (improper tagging and other issues). To avoid this issue our approach is to start with understanding the business goals and then make recommendation for the tool and configuration of the tool.
2. Acting on the findings - One of the major issues is acting on the finding. Once we make recommendations many customers can’t take any actions on them. Why? Because of organizational structure. IT, who is responsible for making the changes to pages, reports elsewhere. Marketing can ask for changes but won't get them because IT has other priorities.
How should this be fixed:
1. Change in organizational structure
2. Help the whole organization know the impact of these changes, show the impact these changes will have on the bottom line.
These are not the only pain points but are the most common that come up almost all of the time. In future I will write about all of the different issues (pain points) that I have come across, so stay tuned.
Sunday, December 03, 2006
Point #1, Avinash states that we should disregard those users who view only a page or stays on the site for less than 10 seconds when calculating the real conversion rate opportunities. In my opinion not everybody who bounces (views 1 page or views the site for less than 10 seconds) can be discounted. These users do present a fair change of conversion. Let me show you why I think that’s the case.
Let’s take an example of a visitor who searches on a keyword on a search engine and lands on the site, three things can happen before leading to the bounce.
1. User spends 5 mins reading the page and then leaves the site. (Assume this is a content site and we are collecting email addresses) Maybe the user will come back maybe he/she won’t.
2. User is so lost when he/she land on the site that she leaves (site is so disorganized or the landing page content or value proposition does not matches with the keyword he/she searched – even though site might very well have that content or product somewhere else)
3. User lands on the site which is not worth his/her time and leaves.– Wrong Site.
In the above 3 scenarios your can only discount number 3 but not number 1 and 2, 1 and 2 provide an opportunity to convert. (Note: Scenario No 3 is also worth looking at; not from conversion point of view but why and how did users land on your site when it was a wrong site for him/her).
Number 1 shows you that user was indeed interested in the site and can possibly sign up for email newsletters had your page provided him proper links or path to conversions. Discounting this user is a big mistake, considering how deeply linked sites are these days and users have lot of information on their finger tips. Users generally won’t spend time to find a path to get converted unless you can convince them, but these user provide an opportunity. This is especially true for lead generation sites, where visitors come to read something specific and they might read only one page in their visit but are valid conversion opportunity.
Number 2 shows that your site’s landing page was not well optimized to lead user to a conversion, you can’t discount the users and ignore this fact. This will be huge opportunity lost.
Note: Time spent on the site is calculated by time lapsed between 2 page views, so when a user views only one page they are automatically excluded from this calculations (not sure if there are tools that can calculate actual time spent even when a user views one page). So don’t discount users with one page views, think about why they only viewed one page what you could have done better to convert them
Regarding customer intent (Point # 3) Avinash says that “One of the biggest mistakes business make is thinking that every visitor to the website is fair game, conversion fodder”. Again, in my view there is always an opportunity to convert visitors to customers. In brick and mortar case I have seen my dad (who owned a retail store in India) converting those people who stopped by in front of his shop for 10 seconds or less or just stopped by to say hello or talk about a product they bought earlier or even to complain about a product they bought in past. If he had taken this approach of deciding that it was not worth his time to pursue those people with no intent of buying then he would have lost a lot of revenue and long term customers. For site which sells lot of retail items this is very true. For example a user comes to the site to gain some information about a product he/she bought in past, say information on how to setup up email on the cell phone, with no intention of buying a new cell phone or an accessory, this visitor can still be converted if the value proposition is there or site does a good job of selling. You can’t discount this fact.
Another thing that should be looked at, in calculating real conversion opportunity, is the effect of cookie deletion on visitor conversion. Cookie deletion inflates the number of visitors and hence your conversion potential. For example, a user comes to the site every day views few pages at the end of the each day deletes the cookies. This user will appear as 30 unique users when you look at the whole month but as you know this is only user. So you only have one conversion opportunity when you use visitor as your denominator.
You don’t have to agree with everything I said but let me know what you think.
Tuesday, November 28, 2006
“I am very interested in Web Analytics fields”, “It is a very interesting field, but how do I start a career in Web Analytics?” I have come across this question (and variations of it) a lot. So I thought, why not write an article to help people who are looking for a career in this 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.
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
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 having diverse educational background.
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 sideline and just read message, 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, this is the best resource for all levels of web analyst especially for those who are just starting. Other books that I recommend are also by Eric Peterson.
1. Web Site Measurement
2. The Big Book of KPIs
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. (I am one of the associate instructors for this course). This course is offered in partnership with Web Analytics Association (WAA)
Reading blogs, articles and whitepapers is another way to expand your knowledge.
Most of the blogs are of advanced nature, so I would recommend you familiarize yourself with the Web Analytics field (see above) before reading these blog. Two of the blogs that I recommend are Occam’s Razor by Avinash Kaushik and Web Analytics Demystified by Eric Peterson, the author of the books mentioned above. Both of these blogs have a list of lots of other blogs on Web Analytics. The more you read the better you will understand this field.
Gaining Practical Experience
Google Analytics (http://www.google.com/analytics) has made it real 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 Paid, while you learn
There are several companies who are looking for entry level analyst. You don’t need any experience, all you need is desire to learn and grow. They will hire you, train you and provide the support to help you grow in this position.
Since you have read this article in it’s entirely, I am convinced you have the desire. Now let me see your passion. Go ahead and email me your resume if you want to get paid while you learn.
Note: I just added a list of books on Web Analytics, you can check them out at Web Analytics Bookstore
Sunday, November 19, 2006
Anyway, even if they are reporting their reach in terms of numbers of behaviors, then how are they counting behaviors? The way I think of behaviors, Revenue Science is underreporting reporting the number of behaviors they reach everyday, it has to be way more than 1 Billion. Let me show you how a single visitor can exhibit over a Billion behaviors.
What is an online behavior?
Every single action that a user takes on the site determines user’s behavior. Following are some of the different elements that determine the behaviors of users online
1. Every page view
2. Number of minutes on a page
3. Path taken
4. Links/Ads clicked
5. Scrolling on the page
6. Referring Sites
7. Each second in the visit
8. Each visit
9. Total Visits
10. Total Page views
11. each Product viewed
12. Each cart abandoned
13. Each step of the funnel completed/abandoned
and the list goes on…...
Let’s take a site with 30 pages. A single user visits all 30 pages. So how many behaviors has this user exhibited? According to my calculation, way over 1 Billion.
How do you calculate online behaviors? (I am only going to count pages viewed to count behaviors)
Each page view by itself is a unique behavior; so this user has exhibited 30 behaviors by viewing all 30 pages.
Combination of pages 1 and 2 is a unique behavior too, that is one more behavior, so total is now 31.
Every combination of these 30 pages will be a unique behavior exhibited by this user.
So how many combinations of 30 pages exist? I am not going to go into details of calculus but show you the formulas here
nCk = The number of combinations of n things taken k at a time
The sum of all the combinations of n distinct things is 2n.
nC0 + nC1 + nC2 + . . . + nCn = 2n
We won’t count any combination with 0 page views (i.e. the user never showed up on the site) so in our example above
nC1 + nC2 + . . . + nCn = 2n - nC0
i.e.nC1 + nC2 + . . . + nCn = 2n - 1
So, combinations of 30 different page views (behaviors) = 230 - 1
That comes to 1, 073, 741, 823 Behaviors. That comes to 1, 073, 741, 823 Behaviors. (http://www.google.com/search?hl=en&lr=&q=2**30+-1&btnG=Search)
Yes over a billion behaviors exhibited by 1 user viewing 30 pages.
If I take various other elements that define behavior (see above) than you don’t even need 30 pages to reach 1 Billion behaviors.
Behaviors motivate but people read your content and buy your products. In my opinion, it is not about how many behaviors a BT vendor can reach, it is about how many customers they can reach.
Sunday, November 05, 2006
Since the beginning of the internet the homepage of the site has been considered the most important page of the site. Till recently, that truly was the case. Visitors entered the site mainly through the homepage; it was the main entry to the site. Cool looking homepage was considered the key to the success of the site.
Visitor’s behavior has changed; homepage is no longer the only entry page into the site. Visitors enter from all sort of different pages and not just the homepage. Homepage is still an important page but not as critical as you might think. Look into the data provided by your web analytics tool before you start changing the homepage in a hope that once the page looks cool everything will be fine. Pretty simple process but a lot of marketers ignore it.
This customer's decision to redesign their site was made without looking into the data their web analytics tool was providing. A simple analysis showed them that only 14% of the visitors were entering the site from home page, and only 24% of the visitors actually saw the home during their visit. They were going to loose an opportunity to convert 76% of the visitors by putting the promotions and the main lead generation link just on the home page.
These days users don't have time to look around. Days of web surfing are long over. Users are generally looking for some specific information/product and once they find it (via one of the factors listed below), they come to the site, get the content/product etc from the site and are gone from the site without checking the homepage.
Below are some of the factors that cause users to bypass the home page.
1. Search Engines – Visitors, now more than ever, rely on the search engines to find the sites or pages which are relevant to what they are looking for. Search engines link into internal pages not just home page, at least that’s the case for well search engine optimized sites.
2. Viral Marketing – Friends referring friends to internal page (pages other than homepage) of the site.
3. Bookmarks – Visitors are book marking the internal pages which have content/products which interest them and then use these bookmarks to visit the site.
4. Deep linking on the other sites – Other sites, blogs etc. are linking to the internal pages with content/products relevant to their content.
5. Online Advertising – Online Advertising (and offline too) drive users to landing pages which are generally internal pages.
I am not saying that you should not have a nice homepage (people do judge the book by the cover, if the see the cover) all I am saying is your redesign should be based on facts and not just gut feel. Focusing just on the homepage is not a smart idea. Your critical promotions and conversion processes should be easily accessible from the top traffic and entry pages, if not all the pages.
As always, I would like to hear your point of view and the stories that you might be willing to share.
Monday, October 16, 2006
Recently I came across a company who invested a lot of effort (time and money) in behavioral ad targeting. At the end, they were not happy with the results and were wondering if it was worth continuing with this effort. Further discussion revealed that they did not get enough clicks on their behaviorally targeted ads and hence their discontent. Since they were paying for clicks only, money was not an issue, but the clicks on behaviorally targeted ads were not enough to justify the efforts they were putting.
So what happened, isn’t behavioral ad targeting suppose to provide you more click-throughs?
The problem was that their behavioral targeting vendor showed them great case study but did not set the realistic expectation about what this customer should expect. I am sure; there are a lot of companies in the same boat. So, in this article, I will explain how to set realistic expectations about what to expect in terms of clicks on your behaviorally targeted ads.
Let's take an example of a Cruise site which sells cruises packages for travel around the world.
This site participates in Behavioral Ad network to target visitors who had shown interest in cruise package to Alaska when they visited their site but wandered-off without making a purchase. These visitors will be targeted when they arrive at other publisher sites participating in the behavioral ad network.
Here are some stats on this site and target audience (visitors)
Traffic: Let’s assume they get 1,000,000 visitors a month (This is for illustration purposes and you can use your own multiplier to see the effect on your site).
Target: Visitors who viewed details and prices about cruises to Alaska. Let’s assume this customer base is about 250,000, 25% of the total site visitors. (Most of the segments I have seen actually are way less than 25% of the total visitors)
Behavioral Ad Network Reach: For this example, the network, the site is participating in, only reaches 30% of the internet population. Let's assume the site’s visitors are a good sample of the total internet population.
By participating in this network the customer will able to reach 75,000 visitors (30% of 250,000), assuming you were able to get impressions on all the sites participating in the network.
Note: This is a big assumption because even if you network reaches 90% of the internet population you will not reach 90% of your target audience. You will be competing with other advertisers (both participating in this network as well as the advertisers directly advertising on the publisher’s site) for the available inventory. Most of the network do not get premium inventory, what they get is Run-of –Site or Run-of-Network inventory which further reduces the reach.
Click-Through Rate: Assuming the click through rate without targeting is .5%. With behavioral ad targeting it is expected to increase to 1.0% (100% increase, this has yet to be proven the % increase is all over the board and one study even said that click through rate declined)
So let's say you show 75,000 impressions to these 75,000 visitors (1 impression each). So the clicks you get will be 750 visitors (1% of 75,000)
So For every 1,000,000 visitors on your site you can only expect 750 visitors from Behavioral Ad Targeting (.075% of your total site visitors, not a huge number).
No let's assume your network reached 80% of internet population. In this case you will be able to reach 200,000 (80% of 250,000) visitors and with 1% click through rate you can expect 2,000 clicks (167% increase). 0.2% of your total site visits, still not a huge number but still better than a network with 30% reach.
By choosing the right network marketer would be able get more clicks for the same efforts.
Your Behavioral Ad Networks (vendor) should have a way to forecast the clicks you can expect. Talk to your account representative; get all the facts, talk to their customer (not just read case studies) so that you have realistic expectation about the clicks. If they can’t provide you this information then time to move on to a new vendor.
Monday, September 18, 2006
To measure online campaigns you assign a unique campaign identifier at the end of the landing url and then use your analytics tool to see how many people responded to the end and then track them all the way to end conversion. It is easy (sort of) to calculate your Return on Investment on online campaigns.
You can use the same method to track offline campaigns, print, in-store display ads, billboards etc.
Here is how it works
1. Create a campaign tracking code(s) to track this campaign just like you do in online campaign
2. Create a easy to remember unique URL e.g. http://www.SeattleIndian.com/saveondining
3. The URL created in step 2 above redirect the users to actual landing page passing the campaign variables.
4. When a user arrives on page created in step 2, the user will be redirect as in step 3 and it will appear in the web analytics tool as the user is coming from a campaign.
5. Add the URL created in step 2 to your print advertising and you are done.
I will be using Google analytics (http://www.google.com/analytics) for this example
1. You have a campaign called “Save on Dining” running as a Half page color in local newspaper
Your campaign variables are
2. Create a easy to remember unique Vanity URL e.g. http://www.SeattleIndian.com/saveondining
3. The URL created in step 2 above redirect the users to actual landing page passing the campaign variables.
4. When a user arrives on http://www.SeattleIndian.com/saveondining will be redirect to http://www.seattleIndian.com/dining.asp?utm_source=Newspaper&utm_medium=Print
Note: You can also set some variables on the Vanity URL web analytics tracking code instead of redirecting to a new URL.
Add the URL created in step 2 to your print advertising and you are done.
So why did I write this article? Well, there are two reasons why I decided to write this article today.
1. I have been involved with tracking campaigns for a fortune 50 company and this topic has come several times. So I had to write this one day.
2. I just read an article by Kevin Newcomb (http://www.clickz.com/showPage.html?page=3623461) about how one company successfully tracked offline campaigns so I thought this is a good time to write it so that users not only know that it can be done but how it can be done.
As always, I would like to hear your experiences with offline campaign tracking.
Saturday, September 09, 2006
Behavioral Targeting (BT) is the ability to target users based on their behavior on the internet. Most commonly it used to target online ads but the technique can be very well used to target products and content.
Behavioral (Ad) Targeting promises to precisely target the audience that matter most. Hit the users with the right message, a message that they care about. It is all about audience.
So How Behavioral Targeting work?
Users are segmented based on the content they view or actions they take on a site(s) and then are targeted with a message (ad) relevant to that segment.
There are two ways behavioral targeting has been deployed
1. On-Site Targeting
2. Network Targeting
The user are segmented based on content views or actions on one site and then are targeted on the site itself.
For example: Users who view 5 or more pages related to auto and view online interest rates page they can be classified as "In Market Auto Buyer"**. Once classified these users can be then targeted with messages from those advertisers who want to reach "In Market Auto Buyers".
The user are segmented based on content views or actions on one site (usually advertisers) and then are targeted where ever they go on sites participating in the behavioral ad network.
For Example: A user views pages 5 or more related to Alaska tours on a travel site, this and then leaves the site without buying. This user is segmented as "Alaska Tour Buyer"**. A very valuable segment for the travel site. This travel site can then tap into the network and share their segment (based on cookie - more on this later) with others in the network and then target them with relevant message to bring them back to the site and convert the sale.
**(How users are segment and what they are called is totally dependent on how marketer wants to define their segment. There is no industry standard, however every now and then I have seen push for creating industry standard way of creating these segment. In my opinion it is too hard to create standards for creating segment as it totally depends on each individual business).
Cool. This is great so should I go ahead and plunge into Behavioral Targeting?
Wait, before you jump into it.
Here are the problem with BT:
1. Most of the publisher sell their inventory based on pre-defined main categories. To integrate BT into their main categories is a challenger for the sales people. The only way big publisher integrate BT into the mix is by selling ROS (Run of site) inventory. Problem is that ROS does not have enough reach.
2. Big publishers do not have any incentive to participate in the network. Again they sell out everything without making the selling process too complicated so why get into this.
3. Because of 2 above most of the networks (covered below) do not have enough reach. Lack of reach makes the opportunity cost too high for advertisers. Why would waste your time and extra money to reach 5 customers. You have a better chance of reaching them without it. I am not saying BT networks should not be used. My point is that tap into a network which has a wider reach.
So who are the major players in this field and have biggest network
There are several players in this field. Since I don't have information on their network, I am not covering them here. I will update this I get more information.
Some of the players I know are
1. Tacoda Systems
2. Revenue Science
3. MSN (just announced their BT Product)
4. I think Yahoo is building one
5. Google is not going to stay behind
I strongly believe that smaller players will vanish (go under or sold) since they wouldn't be able to compete with the reach from likes of MSN, Google and Yahoo.
The real success of the BT is in the "NETWORK".
Feel free to comment on this or contact me if you need any specific information.
Next: I will tell you how you can do On-Site targeting without spending money with one of the BT companies.
Friday, August 25, 2006
Below are some of the things to look at to find out why your traffic is down. Some reasons are within you control (stop freaking start working) some out of your control. Some might be very obvious and some might not.
- Seasonal Impact - Do year over year comparison and see how the traffic pattern was last year.
- May be overall traffic is down even for your competitors (do comparison at http://www.alexaholic.com/ )
- Has a new competitor entered the space? How is their traffic?
- Traffic drivers - How was the traffic from these sources?
Campaigns (Banners, Search, Emails) - Did they ran as normal.
Email - Did anything change there? Did you send out your regular emails, newsletters?
Search - Did you change anything here, has search engines changed their algorithm.
Search - Did you change your site? Meta Tags? Content?
Affiliates - Has any affiliates changed their site.
- Environmental Factors - How is the weather in the geographical region where you have most visitors from? Nice weather can keep people outdoors, resulting in lower traffic.
- Was there any site outage?
- Have you made change to your web analytics tool configuration? If yes, investigate what those were? Problems in filter could be filtering out a lot of traffic.
This is a work in progress. Feel free to send me other things that you might look for.
Monday, August 07, 2006
Know of a book I am missing send the name to me. I have not read all of these but complied this list of of Eric Peterson's Yahoo Group postings.
Web Analytics Demystified by Eric T Peterson http://www.webanalyticsdemystified.com/about_wad.asp
Web Site Measurement Hacks by Eric T Peterson http://tinyurl.com/mzk6r
The Big Book of Key Performance Indicators" by Eric T Peterson http://www.webanalyticsdemystified.com/about_kpi_book.asp
Web Metrics: Proven Methods for Measuring Web Site Success" by Jim Sterne http://tinyurl.com/mptuj/n/n
Drilling down: Turning Customer Data into Profits with a Spreadsheet" by Jim Novo
Measuring the Success of Your Website: A Customer-centric Approach to Website Management" by Hurol Inan
Search Analytics: A Guide to Analyzing and Optimising Website Search Engines" by Hurol Inan
E-Metrics: Business Metrics For The New Economy" by Matt Cutler & Jim Sternehttp://www.targeting.com/emetrics.pdf
Friday, August 04, 2006
Google Analytics is free web analytics tool (ASP) from Google. You can get a free account at
Google Analytics Official Help Links
Help Center: http://www.google.com/support/analytics
Help Group: http://www.google.com/analytics/analyticshelp
Google Blog: http://analytics.blogspot.com/
Conversion University: http://www.google.com/analytics/conversionuniversity.html
ROI Revolution: http://www.roirevolution.com/blog
this just in: http://cutroni.com/blog/
GA Experts: http://www.ga-experts.co.uk/blog/
More coming soon...
I would love to hear your tips and tricks