5 Web Analytics Misconceptions

There are several misconceptions in web analytics (created by some author/bloggers/experts) though many others have tried to clarify them from time to time but they keep reappearing. I recently had a conversation with someone who was so much in love with one of the misunderstood metrics, listed below, that it prompted me to write this blog post. So without much delay, here are the most common five misconceptions that I come across all the time:

  1. More Page Views are good – Unless you are an ad supported site that sell advertising via CPM (cost per thousand impressions) more page views might mean that the visitors are lost on your site and can’t find what they are looking for. More pages views/visit could indicate issue with your site navigation. For effective analysis, set your baseline and then watch for significant deviations (up or down) from the baseline.
  2. All that bounces is bad – I have written 2 detailed posts showing why all that bounces is not really bad. Bounce rate is one metrics that people overly obsess with. Keep in mind all bounces are not bad. The things that cause high bounce rate are:
    1. Links to external sites that you want visitors to click
    2. Ads on your site take visitors out of your site
    3. Returning visits might bounce because they might come to your site to read your daily/ weekly/monthly update
    4. Visits that are for a specific reason e.g. find your phone number
  3. Focus on reducing the bounce rate and everything will be ok– Well that’s the advice many people give without even looking at the other data points and analyzing if reducing the bounce rate will really help you achieve your goal or not. Reducing the bounce rate might not be the most effective way to increase ROI. You should create a monetization model and determine the impact that reducing the bounce rate will have before you start creating different version of a high bounce page to A/B test to reduce your bounce rate. I have seen cases where you won’t get positive ROI even when you reduce the bounce rate to 0%.
  4. Time on site (or page) shows how much time people are spending on the site – As I wrote in my blog post titled Understanding the “Time Spent on the Site” Metrics there are many issues with measuring the actual time spent on the site or a page. One of the main reasons is that the last page that a user views/reads on your site is not counted in this calculation. So if you have a non-ecommerce sites then the chances are that the visitors spend most of their time reading the last page but that page won’t not counted in this metrics and hence your time on page and time on site metrics will be way off. As long as you know that you need to watch the trend instead of using this metrics as a absolute measure of time spent on site then go ahead and use this metrics.
  5. Referring Sites report shows all the traffic sources including campaigns – Well… not really. There are a lot of reasons for the referring source to be lost from the time the visitor clicks on the link to the time they arrive on your site, two big reasons are
    1. Server redirects – This happens a lot with ad serving. Suppose you buy an ad though a 3rd party company who then uses an ad network to place your ads on a publishers site e.g. yahoo, each party does some processing and redirect of its own. In doing all these redirects the referring information is lost or shows one of the sites that does the redirect. For example, you might see atdmt.com showing up in the referring sites which means you were serving ad via Atlas even though the ad might have been served on MSN.com. Many URL shortening services used on twitter also show up as referring domain instead of twitter.
    2. 3rd Party Apps – This is a big issue with Twitter URLs. A lot of twitter users use 3rd party apps and any clicks to your URL posted on twitter from these 3rd party apps will show up as direct traffic.

If you are running a campaign or posting links in social media, blogs, forums etc, make sure to tag them with campaign identifiers so that you can use campaign reports instead of relying on referring sites report.

Read more: 5 Web Analytics Misconceptions – Web Analytics, Behavioral Targeting and Optimization by Anil Batra http://webanalysis.blogspot.com/2010/06/5-web-analytics-misconceptions.html

Adding Twitter Search to Google Analytics

As Twitter increasingly becomes a tool that people use to find information it is really critical for companies to know how people are finding their sites on Twitter. Doing so will allow them to incorporate that learning into future marketing efforts on Twitter as well as other sources (e.g. Google etc).

Since Google Analytics does not yet recognize Twitter search as a search engine like WebTrends or Ominiture does, I will show you how you can do it easily with one line of code.

However, keep in mind this solution only works when the search originates on Twitter (i.e. http://search.twitter.com). 3rd party tools like TweetDeck, those will not be captured in this solution (nor will it, I believe, be captured in WebTrends’ solution). Use this information to understand general search keywords being searched on Twitter but do not get caught up in actual number of visits that your Twitter efforts are driving.

So how do you capture searches conducted in Twitter?

GA provides a few functions to allow you to add your own search engines to the list of search engines that are already tracked by GA.

_addOrganic(newOrganicEngine, newOrganicKeyword)

You simply call this function right after var pageTracker = _gat._getTracker(“UA-XXXXXX-X”); This functions to track any custom search engine.

Twitter uses “q” as the querystring that contains the keyword. So in this case our search engine is search.twitter.com and newOrganicKeyword is the value in query string q

So you code will look like

var pageTracker = _gat._getTracker(“UA-XXXXXX-X”);
pageTracker._addOrganic(“search.twitter.com”, “q”)

(Note: pageTracker._addOrganic(“twitter”, “q”) will also work)

What will the reports look like in Google Analytics?

Note: For some reason I cannot get Twitter search to show up as Twitter in Google Analytics Search Engines report, it shows up as “search”(Maybe it’s a bug in GA? If anybody can provide pointer that will be a great help). However, for now this works fine as long as you know what “search” means in your search engine report. I am playing with filters and if I get that resolved I will post the fix or if you know the fix please email me.

Search Engines Report will show the following:

Drilling down to keywords will show the keywords on Twitter Search.

Looks like Twitter Search brings me repeat visits and more engaged traffic.

Originally Posted at : Adding Twitter Search to Google Analytics – Web Analytics, Behavioral Targeting and Optimization by Anil Batra in Feb. 2009

Getting Started with Web Analytics

This is a beginner guide to getting started with Web Analytics.

The easiest way to get start with web analytics and start tracking your own site or blog is to use Google Analytics.
You will need:

  • A Google Account – to sign up for Google Analytics
  • Access to your sites code to add a small JavaScript code to all the pages.

Sign into http://www.google.com/analytics with your Google Account and follow the instructions on the screen.

Finally you will get to a page that will give a JavaScript code that you will need to copy and paste on the source code of all the pages on your site. If you site has a common header or footer file then just stick the code in that file. If you do not have a common header or footer then you will have to copy this code on every page.


After 24-48 hours you should start to see the data in your reports. You can access your reports by going to http://www.google.com/analytics

3 Roles in Web Analytics

Despite slow economy many companies are hiring web analysts. A quick search on Simplyhired.com, a site that powers the Web Analytics job board on my blog, shows that there are currently 2,007 open positions and indeed.com, another job sites shows over 4800 open positions. That is a huge number.

However, many job seekers I have talked to feel frustrated because most of the jobs have a laundry list of requirements and they don’t feel that they are a right fit for most of these open positions. A lot of “Web Analytics” job openings ask for many of the following:

  • Experience in online marketing
  • Experience in Web Analytics
  • Experience in – Google analytics, Omniture, Webtrends, Coremetrics etc.
  • Experience in implementing Omniture, Google Analytics, WebTrends etc.
  • Experience in A/B and Multivariate testing
  • Experience in Search engine optimization
  • Experience in search engine marketing
  • Experience with SQL
  • Experience in email marketing
  • Experience in Social media

The mismatch in what a company really needs and what they are asking in the job requirements is a cause of frustration on both ends. The issue really stems from lack of understanding of what web analytics is and what role a web analyst need to play in the organization.

Most of the companies looking for a “web analysts” are in one of the following three stages of web analytics staffing

  1. They don’t have any tool but they realize the need and are looking for someone who can help them with “web analytics”.
  2. They just installed Google Analytics or were sold one of the other paid tool but are not getting much value from their web analytics tool. They need an analyst to help them do “web analytics”.
  3. They already have a web analytics tool installed and have a web analytics team. Since the company is now using web analytics to made business decision they need to hire one or more analysts to support the growing demand.

Companies falling in the third stage know what they are doing and usually narrow down the requirements. They are usually clear on what kind of person they are looking for.

Companies who fall in stage 1 and 2 above are the ones who are usually not clear on the role of a “web analyst” and hence create this laundry list of skills. Hiring manger looks at few job openings posted by other to get an idea of what a “web analyst’ should do. She then includes all the buzzwords and sends the requirements to HR or the recruiting company. HR screens the resume and if the keywords shown above do not appear on the resume the resume is rejected. As a result, companies loose several good candidates while candidates loose many good job opportunities.

3 Roles in Web Analytics

If you are a hiring manager, you need to understand and thoroughly evaluate your need before opening the job req. This will help you remove the noise from requirements and find the best candidate for the job. To make your job easier I have categories web analytics work into 3 job roles.

  1. Implementation Specialist/Engineer

    If you are looking to implement a web analytics tool then you will need an Implementation Engineer. Implementation Engineer is usually the one who manages implementation of the web analytics tool and/or maintains ongoing implementation changes. This is a technical role. For this role you will need a person who has experience in implementation of the web analytics tool of you choice (Note: Tool Selection is a complex process and you should hire a 3rd party consulting company to help you with it if you have not already selected the tool). An implementation engineer generally takes the business requirements and converts them into technical requirements for the web development team to implement the code on the pages. Implementation Engineer works closely with “Web Analyst” (described below) web development and QA to ensure that correct data is collected. The right candidate for this role understands how internet technologies work. She needs to have a good grasp of JavaScript (most of the web analytics implementations require JavaScript tagging). She might also need to understand how to integrate various data sources together. For many companies, once the tool is implemented there might not be a daily need to make changes to the tool so it might make more sense to outsource this function to a web analytics vendor, agency building/maintaining your site or a web analytics consulting company instead of hiring a fulltime person.

  2. Reporting Analyst

    Are you doing web reporting or web analytics). If you are looking for someone to pull the data from your web analytics tools or other reporting application then you need to hire a reporting analysts. A lot of the companies confuse “web reporting” with “web analytics”(See my blog post titled Reporting analysts usually understands the interface of the various tools and can pull the data that is required by other stakeholder. A reporting analyst might need to have SQL skills to pull the data from databases. Some organizations might need a person who can make pretty scorecard and charts. For this role, it is good to have a person who has experience with the tools of your choice but don’t make it a deal breaker. If the candidate has worked on any of the web analytics tools then she can usually get trained in other web analytics tools. Determine what other tools do you have and what skills might be required to pull the data from all these tools, that you might need for you reporting and then write the job requirements.

  3. Web Analyst

    This is more of a business role and truly a web analyst’s role. This is a person who can make sense of the web data and drive insights to impact the bottom line. She will provide business requirements to the Implementation Engineer to work on and will use reporting analyst to get the data for analysis. Web analysts are inquisitive and analytical, they question the data to come up with the story that the data is telling. Web Analyst has the ability to understand and analyze various data pieces such as competitive, qualitative, web analytics, social media, financial etc and drive business changes. Web Analyst should also be able to run A/B and Multivariate tests to improve website performance. Depending on the size of your organization and A Web Analysts will not be afraid to stand in front of executives to explain and defend their findings. If you are looking to get actionable recommendation and drive business changes based on web analytics data then you need a Web Analyst.

Hope this will help you in properly wording your job requirements and avoid the frustration of not filling the positio

Read more: 3 Roles in Web Analytics – Web Analytics, Behavioral Targeting and Optimization by Anil Batra http://webanalysis.blogspot.com/2009/08/3-roles-in-web-analytics.html#ixzz1G8UjaJ8W

Skills Required for a Web Analyst – Part II

This is the second part in the two part series of “Skills required for a Web Analyst“. You can read the first part at http://optimizationtoday.com/web-analytics/skills-required-for-a-web-analyst-part-i/.

Here are some of the responses from :

“To me, the Web needs three ingredients: clear business objectives, a communication strategy and the technology to support them. The best web analyst would be savvy and top of the line in all three… which is probably impossible to find. But everyone can rate himself on a virtual scale and see where they stand and where they want to be.”
Stephane Hamel, Web Analytics freelancer and creator of WASP tool

Curiosity and enthusiasm. I know these aren’t skills so much as innate qualities, but I believe if you can go a long way as a web analyst with curiosity and enthusiasm. If you are genuinely interested in what you’re doing, it shows.”
June Dershewitz, Vice President Semphonic

“Analysts work at various levels — the most fundamental level requires some understanding of basic statistics, such as linear regression, seasonality, and simple trend forecasting, combined with an understanding of how the web “works.” In other words, what do we know about browsing behavior, combined with cookie deletion probabilities, web server caching, and other artifacts of the web experience, to put our data into a reasonable context, from which we can draw useful conclusions?

Beyond that basic level, though, is an understanding of organizational politics, and some fluency in “managing your message” so that the analysis you’ve worked hard at creating actually influences managers in making better decisions. Otherwise, you’re just creating “doorstop reports,” that don’t help your organization, and don’t really justify your efforts.”
WDave Rhee, Analytics Country Manager, Germany, OX2 / LBi Group

“I think 2 of the biggest skills required to be a web analyst is to be able combine technical skills along with business/marketing skills in order to understand your customer’s goals and provide value to them. Secondly you have to be passionate about what you do, if you don’t like what you’re doing you will not give 100% effort.”
Manoj Jasra, Director of Technology at Enquiro

Skills Required for a Web Analyst – Part I

Web Analytics is one of the hottest career fields and it is becoming very hard to find experienced web analyst. One way to counter this shortage is to hire the people with the skills them could eventually make them a great Web Analysts and then train them in the Web Analytics.

So what are the skills that you should look for when hiring people who don’t have years of Web Analytics experience? This was one of the questions I asked when I interviewed a few of Web Analytics professionals (Note: these interviews were conducted about 3 years ago but the answers are still relevant). In this 2 part series I have compiled their responses, which should help you in understand what skills you should look for when hiring a web analyst. If you are trying to enter or switch to the field of Web Analytics then this should give you an idea of what skills you need.

Here are some of the responses:

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.
Trinadh Rao, Country Manager Web Analytics Association, India Web Analytics Manager at Franklin Templeton

A great deal of tenacity and, being sharp enough to make the connection between right and left brain items.
Daniel Shields, Web Analyst at CableOrganizer.com

In my opinion a web analyst needs to understand the data and site dynamics of their particular website. Should have decent technical knowledge, good customer focus and attention to details. Should intuitively pick insight from an otherwise burgeoning deluge of data.
Jaisiri Chetty, Asst. Manager (Insight), Tesco.com

In terms of education any advanced course in statistics should work as a passport to Web Analyst job. He/She should be genuinely good at collating different industry information.
Apurba Sen, Product Manager at Rediff.com India Ltd

A sense of abstract thinking and Art (seeing the whole picture at once) is important for Web Analytics work.

I think what a lot of companies are asking for has nothing much to do with what is needed to be truly effective. In order to be effective as a Web Analyst you don’t usually need a degree in statistics or be trained as a rocket scientist (yet that is what all the job descriptions I’ve seen ask for – like a big long laundry list) you need the ability to understand what some one needs to measure, what the goals are, and a technical mastery to the tools, the web analytics platforms being used. The least important thing is to know the tools beforehand – because anyone can learn them fairly quickly who is not brain dead (yet this is precisely what most interviewers ask for first – knowledge of the tools)

The most important thing, in any job, is being effective in the job you’ll be hired for –being trusted, that you can deliver what say you can deliver on…and a lot of that is based on trust, on conveying confidence, conveying authority. The technical part is more like the icing the cake, as far as I’m concerned. I go for rapport, gut feeling, intuitive knowledge and visualization of what my clients want and need, and even what they don’t know to ask for, but which they still need…I try to give them…and this is not just at IBM, but all my work is done like this.

You know you’re effective when you’re allowed to work on the “big problems” in your organization. It’s a feedback loop. You need to get trust of higher ups so they’ll let you get the relevant experience that actually makes you valuable in the marketplace (so that you can “move up”). The paradox is, believing in your self first, is necessary in order to get any kind of trust and buy in so that you’ll be allowed to work on the big stuff.

You also need people to like you – and for some people, that’s easier to achieve than others. I won’t say that people that are disliked are not effective – they can be also, but they’re probably miserable and less effective than if they were liked.

But none of these skills is actually what is asked of you in an interview – yet some interviewers will make note of them, nonetheless and the one’s that do are the one’s to work for.
Marshall Sponder, Blogger at Webmetricsguru.com

More to come in part II.

Original Post: Skills required for a Web Analyst – Part I – Web Analytics, Behavioral Targeting and Optimization by Anil Batra http://webanalysis.blogspot.com/2008/02/skills-required-for-web-analyst.html#ixzz1FkEREDmq

Starting a 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. This article is 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
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 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.

Books that you should read:

1. Web Analytics Demystified
2. Web Analytics An Hour A Day
3. The Big Book of KPIs
4. Web Site Measurement
5. Advanced Web Metrics with Google Analytics
6. Always Be Testing: The Complete Guide to Google Website Optimizer

Note: I just added a list of books on Web Analytics, you can check them out at Web Analytics Bookstore

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. Also check out Post Rank for a list of Web Analytics Blogs.

Twitter: At twitter, you can find a list of web Analytics folks and topics by searching for #measure.  There is also a list of twitter users in the web analytics community at http://webanalysis.blogspot.com/2008/12/my-view-on-twitter.html.

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).

Another great resources is Analysis Exchange. Analysis Exchange has created a training ground for new analysts all around the world.

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.

What is Abandonment Rate

Drop off or abandonment rate measures the number of visits/visitors who left a conversion process (funnel) without completing it.  Any  process with 2 or more actions on the site can be considered a conversion process, what you define as a conversion depends on the purpose of your site and your business objective. Some of the commonly used conversion funnel are shopping cart, newsletter signup and document downloads.

Abandonment rate helps identify the steps in the funnel that are causing the users to drop off.  Conducting analysis of those steps will help us take necessary steps to minimize the drop offs and optimize the conversions.
There are 2 ways to calculate Drop off rate and each of them provide the data in slightly different ways. Both of them are correct ways to calculate.

  1. Total Drop off from First Step = (Visits to the current Conversion Step-Visits to the First Conversion Step)/Visits to the First Conversion Step.  Let’s assume that we want to use Product as the first step of the conversion process and that step gets 10,000 visits.  But of those 10,000 only 7,000 continued to the next step of adding the product to shopping cart.  Then in the next step only 200 out of total 10,00 that started the process continue to Registration form and finally 1,200 out of 10,000 got to the final confirmation page. This means we saw 30% abandonment between Step 1 and Step 2  ie. (7,000 -10,000)/1000.  Abandonment was 80%  from Step 1 to Step 3 (2,000 – 10,000)/10,000. Final abandonment rate was 88% (1,200 – 10,000)/10,000.  Though this calculation gives us a good idea of final drop off rate and conversion it is a little hard to understand.
  2. Step Abandonment Rate = (Visits to the current Conversion Step-Visits to the previous Conversion Step)/Visits to the Previous conversion Step. This calculates the abandonment rate between the two immediate steps.  Unlike the previous calculation it provides a better idea of abandonment as visitors move from one step to the other. As shown below, again out first step had 10,00 visits, of those 7,000 continued to next step resulting in an abandonment of 30% (same calculation as above).  However when the visitor moves from step 2 to step 3, our abandonment rate now is is 71% (2,000-7,000)/7000. The final step then shows an abandonment rate of 40% (1,200 – 2,000)/2,000 Using such a calculation you can easliy see which steps are causing the biggest drops, in this case it is step 2 and step 3, people are adding products to cart but not clicking on the checkout button to go to registration form.  Why? Maybe checkout button is not visible? Maybe it is the name of the button? Whatever the reason is, you have the data to show where the problem is and gives you an idea on what needs to be fixed.

Contributed by Rohan Kapoor based on his original post at Funnel Drop Off/Abandonment Rate

  1. the above formulas, it looks like the first one seems better in terms of Funnel visualization but personally I like the second formula better. I say this because in the second funnel we are only considering the respective Conversion steps in the calculation and not Step 1 (Homepage) because Step 1 is entirely a separate user experience. According to me the Drop off rate should be calculated based on the 2 Conversion Steps as they are independent of the user acquaintance on the other pages of a Funnel. These 2 pages alone can determine how we can improve the conversion rate at each step as these are not based on the Homepage experience. For e.g. The Registration form design and involvement is totally different than what it is on the Homepage. I hope you like this post and would really appreciate if you can share your opinions.

Contributed by Rohan Kapoor based on his original post at Funnel Drop Off/Abandonment Rate

How to Measure Online Advertising Success

I am amazed at how companies spend millions of dollars on online advertising but none to actually measure if it was successful or not. I have come across several companies in past few years so thought I will share my 10 step process to measuring the success and ultimately improving the ROI.

Below are two eye-opener real life examples that will show why I thought this was a subject that I should blog about:

  • A customer spent 8 million on a huge online campaign but had not clue weather they were getting their money’s worth or not. All they got was banner impressions and initial click through rate (CTR ) from their agency. This initial CTR was in line with what their agency had expected so they were contended with the results. As far as measuring beyond the initial CTR they had no idea. Their answer was that we do not sell anything so we can not see if this is generating money or not, all we need to do it generate brand awareness. Well were they generating brand awareness? In few minutes we were able to see that that they had 90% bounce rate (yes they had WA tool implemented but were not looking at it, yah I know what you are thinking). That is 90% of the money down the drain. It is true that everybody who gets to the site has been exposed to the brand but is that enough? 90% bounce rate was pretty substantial considering that initial click through was close to 1%. I don’t think they were able to generate brand awareness.
  • Another customer spent about 4 million on an online campaign but was very stingy when it came to using web analytics tool on their site to measure the success. Not sure if the marketing manger was not comfortable with the result that she would get or just did not consider it worthwhile to measure because she had extra money to spend.

Does this sound familiar? If yes and you are marketing manger or marketing exec, I would strongly suggest putting some money aside for measuring your campaigns performance beyond the initial CTR. You will be able to learn a lot more about how your campaign is performing and improve your ROI by simple A/B testing. If you are an analyst then please work with the marketing department and sell them the benefits of measurement.
Below are my 10 steps for measuring the online advertising success, nothing fancy, a simple straight forward process that will improve your bottom line.

  1. Determine the goals and objectives of your campaign. – Knowing why you are running the campaign is first and foremost step. Unless you know why you are running the campaign you will never know if was successful or not. Marketing manager and web analyst should be in sync on this. Infect all the stakeholder should be on the same page. A clear understanding of the goals helps everybody focus on same things.
  2. Determine success criteria and KPI’s for your online campaign. – Once you know the goals of the campaign, next step is to determine the Key Performance Indicators (KPIs) of the campaigns. These key metrics will allow you to see how the campaign is performing. If you already have a baseline measures from your previous campaign then you can compare your metrics against them or if you don’t have one then within weeks or a months (depending on the duration of the campaign) you should be able to develop one.
  3. Create a campaign attributes framework – It is very important to build a campaign attributes framework from the beginning. Deciding what attributes to measure the campaign against early on will make sure you capture them from the beginning. Also upfront thinking will allow you to get a buy-in from agency, as most likely they will be responsible to for providing you with all the campaign attributes. Some examples of the attributes are placement, creative, message, publisher.
  4. Implement proper web analytics tracking code on your site and landing page(s) – After you have the framework in place and know which metrics to capture, next step is to get together with you implementation team to implement proper tracking code on the landing pages and site. As you already know if the tracking codes are not implemented properly you will not be able to track your campaigns. I suggest running a test before you go live, so that you can resolve any issues upfront else it will be too late.
  5. Configure web analytics tool to measure your KPIs – Another area to pay close attention to is tool configuration. Too many times I have seen disconnect where KPIs are determined but the tool configuration is so messed up that you can not measure anything. Determine what reports you will need and how they will be configured. Work with your implementation team; make sure they understand the goals and what you are trying to measure. I was recently involved with a campaign measurement configuration where we had to reanalyze the data because the implementation guy messed up one configuration. Luckily point no.4 above was correct so all the data was there it just had to be reanalyzed. Pay extra attention.
  6. Tip – Configuration of the tools (reports) should be such that you can compare overall user (or user driven organically) with those driven by the campaign. I have found that this kind of insight helps you better understand your visitors and determine where you should be spending your money and effort. The more you can segment your user base the better insight you will get.

  7. Build a scorecard or dashboard that will allow stakeholders to focus on the KPIs – The temptation by stakeholders is to look at every single data point weather they understand the impact of those or not. I have been in meetings were they will try to argue on things that don’t even matter. Why? Because they had access to the data and like to argue. The scorecard or dashboard should allow them to focus on things that matter instead of every data element that a web analytics tool can provide.
  8. Tag all your ads with appropriate campaign identifiers – You have determined the KPIs, setup the reports and a nice scorecard is waiting for the data and analysis, but guess what? The agency did not add the proper campaign identifier. All your efforts are down the drain. The problem occurs because people setting up the campaigns in ad servers have zillion other things to worry about and if they are not ingrained in the process they will forget or won’t give due attention to campaign tracking. As I mentioned above in point number 3, if you create the attribute framework and involve the agency then this should not be an issue.
  9. Analyze the data in few hours of launching the campaign, fill the scorecard and learn from the data. Few hours might be too early to learning anything meaningful (depending on the magnitude of the campaign) but can show you if something is really screwed up. Tune as necessary.
  10. Periodic reporting and analysis (will depended on the length of campaign but start with daily then weekly/monthly). – Periodic reporting and analysis is an important aspect of this process, this is where you will actually know if the campaign is achieving its goals or are there things that should be changed. Don’t stop at reporting only analyze the data (see my article titled Are you doing Web Reporting or Web Analytics. Provide actionable recommendations. Provide your analysis back to the stakeholder. Discuss them, debate them and determine what to test. (I am not going to go in detail on A/B or multivariate testing in this article)
  11. Put this process in place and share with all the stakeholders. Buy-in from all the stakeholders is necessary for any process to work. Put some timelines so that key stakeholder can be involved in the process in timely fashion.

The above process will help you measure the true success of your campaigns. Learn from the data and optimize as necessary. Remember there is always room some for incremental improvement. You will be amazed how you can improve your ROI by following above 10 steps.

Contributed by Anil Batra: 10 steps for measuring online advertising success – Web Analytics, Behavioral Targeting and Optimization by Anil Batra http://webanalysis.blogspot.com/2007/06/10-steps-for-measuring-online.html#ixzz1FNi3zbHH

Referring Domains Demystified – Part I

A while ago a user of the yahoo webanalytics group asked why their own domain was showing up as the number one referring domain. Interestingly enough another person asked me the same question the same day. I have also seen this question pop up several times in past so I thought why not clarify for everybody.

Note: I know the images on this post are very small. I am trying to figure out how I can make them bigger but had no luck so far. As soon as I figure out hot to fix it, I will update this post with better pictures

This will be a two part series. In this part I will explain how Web Analytics tools report the referring domains and pages. I think this is a very critical aspect before we dig into why your domain shows up the referrer in referring domain report. Part II will explain why your own referrer shows up in the referring domain.

So let’ begin with understanding what a log file contains:.

What is contained in a log file:: (Note: Even tag based solutions generate a log file, although they don’t call it a log file. This log file is in the vendor’s proprietary format but for this example you can assume it looks the same as the example below).

Every log file contains a line for each of the server ip, request file (URI stem of the file), with time stamp, a cookie (if one exists), and the referrer (URL of the page that referred user to the current file/page). There are other fields as well but they are not required for this example. The assumption here is that if you are using your server logs and they are in W3C Extended Log File format or you are using the JavaScript solution of your web analytics vendor.

Here is a list of all the fields in W3C Extended Log File Format:

Here is an example of the log file

You can read more about is at: http://www.microsoft.com/technet/prodtechnol/WindowsServer2003/Library/IIS/ffdd7079-47be-4277-921f-7a3a6e610dcb.mspx?mfr=true

Below is a an example of the log file from http://www.seattleindian.com

#Fields: date time c-ip cs-method cs-uri-stem sc-status sc-bytes cs-version cs(User-Agent) cs(Referer) cs(Cookie)
2007-02-21 07:07:30 GET /Seattle/moviepictures/bollywood.asp 200 33453 HTTP/1.1 Mozilla/4.0+(compatible;+MSIE+7.0;+Windows+NT+5.1;+.NET+CLR+1.1.4322;+.NET+CLR+2.0.50727) http://www.seattleindian.com/seattle/ USERID=1234;+RID=3423

In this log file /Seattle/moviepictures/bollywood.asp was the current page that the visitor viewed and it was referred by http://www.seattleindian.com/seattle/. Every log line has the referring page that contains the URL of the page that referred the user to the current page.

Let’s take an example of visits for one visitor. For this example I am only showing 5 fields (s-ip, data, time, URI stem, cs(referrer)

Below is the data for a visitor:

The visit started with a referral from http://www.google.com/?q=seattleindian. The referring domain in your web analytics tool will be Google.com

Let’s assume, this visitor goes on a lunch break leaving the site open in her browser. Come back after an hour (note the time change) and clicks on the home page links, here is how the log file will look like

This constitutes a second visit (assuming a 30 min session time out). The referring page will be http://www.seattleindian.com/seattle/advetise.asp and the referring domain will be SeattleIndian.com for this second visit.

This is one of the reasons why you will see lots of visitors referred by your own site. We will cover other reasons in Part II.

Let’s assume that this user clicks on a link on SeattleIndian.com that takes her to a new domain, SFIndian.com. This domain is identified by a different server ip ( in the log file. Here is the log file for this other domain

Referring page for this visit will be http://www.seattleindian.com/seattle/advetise.asp the referring domain will be Seattleindian.com.

SeattleIndian.com and SFIndian.com are both owned by the same owner who wants to see the combined traffic. Let’s assume you were to combine both the domains in one profile (or reporting suite) and assume they were using the same third party cookie. This is how the combined log file will look like

When your web analytics tool parses these log lines, it will report 2 visits

One visit started at 10:59:50:00 and ended at 10:59:50:04 and the referring domain was Google.com

The second visit started at 11:59:50:05 and ended at 11:59:50:13 on SFIndian.com, the referring domain was SeattleIndian.com

Now let’s take an example where user goes out to another site and then comes back to the site

In this example: The visit started by referring domain being google.com, the user leaves the site from bollywood.asp page and comes back via yahoo search to aboutus.asp page. In this scenario the visit started via google so yahoo will never show up in the referring domains (that is the case with most of the web analytics tool).

Note: If you run a campaign on Google and Yahoo and tag them so that your web analytics tool can track them, then you can configure the tool to show either the last or the first campaign. However; the out-of-the-box referring domain report will only show the first referrer of the visit.

Hope this clarifies some of the confusion around referring domains and pages.

Read more: Referring Domains Demystified – Web Analytics, Behavioral Targeting and Optimization by Anil Batra http://webanalysis.blogspot.com/2007/03/referring-domains-demystified.html#ixzz1EjZaDjXt