SEM Vendor Queries
‘ppc management [vendor name]‘
‘Google recommended bid management tools’
‘[vendor name] competitors’
‘[Vendor 1] [Vendor 2] [Vendor 3] [Vendor 4]‘
‘top SEM firms’

It’s perhaps no surprise that when people responsible for SEM at an advertiser or agency are looking for SEM tools, they go to search, but what is surprising is how little objective content they find. Do a search for any of the above queries, and what you’ll find is that aside from industry-specific blogs such as mine, there are no independent third party reviews on SEM tools to speak of. Not so for more mature technologies such as CRM, databases, industrial supplies and web hosting. Perhaps someone should take this bull by the horns? Anyone? In the meantime, SEM technology buyers should post on forums such as Webmasterworld, SearchEngineWatch and DigitalPointForums asking others what experiences they might’ve had with Vendor X, Y or Z. Also, if you’re going to settle on the right solution, see who is & isn’t advertising on the SEM-related queries you’d expect them to, and keep in mind that irrational SEM bidding by SEM vendors is a sign they probably won’t take good care of your own campaigns.

SEM Sales Rep Queries
‘Chris Zaharias Omniture’
‘Anil Kamath Efficient Frontier’
‘Kevin Ryan Motivity’

No doubt the SEM buyer is trying to learn about the SEM salesperson they have met or will meet, but more importantly, I think SEM buyers want to know that the people they’re speaking to out SEM tools actually know what they’re talking about, participate in industry conferences and discussions, and have enough knowledge of the space for their viewpoints to be worth considering. Over the years I’ve had many, many business relationships start with people coming to me and saying, in effect, ‘I wanted to get your opinion on your and other solutions because I keep seeing your name as I research the space.’ Horn tooting aside, this really means that most of the people representing SEM firms in pre-sales discussions don’t know enough about the space to be trusted as capable of giving accurate assessments of their and other vendors’ solutions. And further, these SEM credibility searches make clear the unfortunate predicament SEM buyers face: there is no objective source of SEM vendor reviews.

SEM Campaign Management Queries
‘history of bid management’ - surprisingly, there are quite a few students of the space, but very little in the way of Wikipedia entries or historical SEM articles. This will change, but only if SEM buffs do their part.
‘Google match types’ - probably the most frequent query I see, reflects consistent industry concern that the match types Google defaults to are not the ones that best serve advertisers.

‘Expanded Broad Match’ - likewise, you never see people searching for Exact or Phrase match ; those match types work the way you’d expect them to, whereas EBM is much more challenging for advertisers to understand.

‘Automatic Match beta’
’search query length’ - I’m seeing more of this query lately, perhaps because people are seeing shrinkage of the long tail in their own campaigns.

‘SEM optimization algorithms’ - not as rare a query as you’d think, shows that buyers know SEM systems’ math is an important consideration.
‘ppc distribution fraud’

‘AdWords account hijacked’ - little has ever been written about what is actually a very regularly occurring PPC-related crime, but hardly a month goes by without someone on an SEM discussion forum claiming that their account was hijacked and new campaigns added that drive traffic to a fly-by-night affiliate program.

Perhaps just as interesting are the SEM queries you don’t see. Why aren’t people searching for
‘best SEM user interface’
‘SEM buyer’s guide’
‘what to look for in SEM solutions’
‘SEM vendor reviews’

Because the majority of SEM solution buyers are first-time buyers, most don’t know what they don’t know and just assume that if they research the vendors that contact them, they’ll do fine. That couldn’t be further from the truth, though, as SEM firms are notoriously good at - lo & behold - using search to identify prospects and will take advantage of novice buyers like you wouldn’t believe. So as an SEM solution buyer, recognize the lack of objective information, talk to as many other buyers as possible, and know the vendors from multiple viewpoints online and off.

That’s really encouraging news, especially in a downturn economy. It shows that marketers are realizing that testing is one of the few strategies that truly enable you to do more with less. Instead of having to invest linear amounts of money to acquire traffic, you can instead invest your ideas and hypotheses to increase conversion and revenue on your site.

I often find that the Q&A is the most interesting and engaging part of a session. It’s the closest we get to understanding what’s really on the minds of marketers today, and what they were hoping to learn from those of us on the panel. One of the questions that stuck in my mind was, “What are good strategies for optimizing social media traffic going to landing pages?”

To be honest, I didn’t have a great answer because I haven’t seen many tests involving social media traffic. Maybe it’s because we don’t have customers who are investing much in social media yet. But I thought about it more on the plane, and then at home, and I think the answer lies in going back to the standard best practices of first defining what your success event will be, and then forming a hypothesis around how to improve it.

For example, if you told me that you ran a content site and wanted to increase page views, I’d say that those who come from social media probably have a higher likelihood of sharing content with friends. My test idea might be to test promoting sharing links higher up on the page, or repeatedly throughout the article. But if you told me that you ran a retail site and you wanted to increase orders, I might try to promote highly-rated products based on the hypothesis that social media visitors place a greater premium on the opinions of peers.

Another question I got was, “Okay, you talk a lot about testing, but how do I test on my site? Do you sell a product?” Yikes, the dreaded vendor pitch trap! I have to admit there was a bit of spastic product naming as I tried to give a very neutral answer. I struggle quite a bit with the balance of giving information and opinions without selling any specific platform, both in places like PubCon and here in this blog. I know that you, as the reader, have come here knowing that it’s a corporate blog run by Omniture, but I also think that you’re more interested in learning about marketing principles and customer insights than about what we’re selling. So I err on the side of avoiding the product conversation completely, to the detriment of those who are already customers. Luckily, my very capable colleague, Brian Hawkins, has begun a blog recently about the ins and outs of Omniture Test&Target, so I would direct readers there for specifics on how to implement many of the examples and case studies we show.

The last question I have is for you, the marketer who attends conferences like PubCon. When you go to a session about landing page optimization, what are you hoping to hear? Are you looking for tactical advice, or customer case studies, or market research findings? Would you rather see a presentation that tells you about the latest marketing strategies and where we see the market going, or the one that lays out 10 easy tips to try on your site today?

Understanding The Products Variable
As its name implies, the primary use of the Products variable is to store the name of the Product for which a website action is taking place.  For example, if you manage a retail website and a visitor views a Product Detail page, you might choose to set a Product View Success Event so you can see how many Product Views your website had each day, week, month, etc…  But if you want to see which Products visitors were looking at, you would need to capture the Product name or a Product ID# in a variable so that you can break down the Product Views Success Event metric by Product.  For those who have read previous posts, your first instinct should be to use an Conversion variable (eVar) to capture the Product Name, since eVars are used to break down Success Events.  While you could certainly do that, Omniture has assumed that many websites will have Products so it has created a special Products variable for this exact purpose.  Therefore, it is recommended that you capture the Product name or ID# in the Products variable (s.products).

“What is the difference between an eVar and the Products variable?” is the next question I get from clients (which is why I am writing this post!).  There are some subtle differences between the Products variable and Conversion Variables (eVars) that you need to understand.  They are as follows:

Multiple Values
The Products variable allows you to pass multiple values to it whereas eVars are limited to one value.  The importance of this cannot be understated.  When a Success Event is set in conjunction with an eVar, whatever value is passed to the eVar gets credit for the Success Event.  But if multiple Products are passed to the Products variable at the same time that a Success Event is set, each value passed to the Products variable will get credit for the Success Event.

Not Persistent
Another way in which the Products Variable is different from Conversion Variables is that it is not persistent, meaning that it does not retain its value from one page to the next.  In this respect, it is somewhat similar to Traffic Variables.  For this reason, if you need to associate multiple Success Events with specific Products, you must set the Products variable every time you set Success Events.  For example, if a visitor adds three products to the shopping cart, you would pass all three to the Products variable when you set the Cart Add Success Event, but if later the visitor purchases one of those products, you need to pass that product to the Products variable again when you set the Purchase Success Event.

Special Parameters
Unlike Conversion Variables, the Products variable allows you to set some special parameters at the same time the Product value is set.  While this can be a bit confusing for newer SiteCatalyst users, I will do my best to explain this in non-technical terms.  To begin, it is important to see the full Products Variable String with all of the potential parameters that can be set:

s.products=[Category1];[Product1];[Quantity];[Total Price];[Incrementor];[Merchandising], [Category2];[Product2]; etc…

The following will describe each:

[Category] - This represents the Product Category.  It is not recommended that you use this parameter due to the fact that it ties the Product to the first Category that it is associated with and it cannot be changed.  There are now better ways to assign Products to Product Categories so this parameter is normally not set and remains primarily for historical purposes.

[Product] - This represents the name or ID of the Product.  If you pass an ID, you can always use Classifications to upload friendly names and roll products into Categories.  If you will be passing a lot of Products at one time, it is recommended that you use ID’s instead of full product names to limit the length of the overall product string so you do not exceed browser character limits.

[Quantity] - When used with the Purchase Success Event, this represents the quantity of the Product being purchased (i.e. visitor is buying two memory cards)

[Total Price] - When used with the Purchase Success Event, this represents the total price for the Product being purchased (i.e. two memory cards total $200 for both)

[Incrementor] - This is an advanced topic which I will be covering in the future, but in general, you can set an Incrementor Success Event such that you manually pass a currency amount or number to it.  For example, you charge $2.50 shipping for a product and want to show that separate from Revenue, you can devote a Success Event to “Shipping Costs” and pass “2.5″ in this part of the Product String to add $2.50 with each purchase.

[Merchandising] - This is another advanced topic which I will cover later, but in general, you can use this parameter to bind Products to different eVar values for each Product instead of tying all Products to one eVar value.  This is often used to capture which product category the visitor used to find the Product.

So in a scenario where a visitor is buying two XYZ memory cards at $100 each, has to pay $2.50 shipping and found the memory card product through the “Sale” section of the site, the Product string might look like this:

s.products=”;XYZ Memory Card;2;200;event18=2.5;eVar12=Sale”

Important Things To Know About The Products Variable
The following are some important things to know about the Products variable:

1. The Products variable has full subrelations so any conversion variable report can be broken down by Products and any Classifications of Products.
2. If the Products variable is set with no corresponding Success Event, the “prodView” Success Event will be used by default.  This will appear in SiteCatalyst as the Product Views metric and only appears in Products reports.  However, if Product Views is something that is important to your organization, it is a best practice to use a Custom Success Event to track Product Views.
3. The Revenue and Unit parameters are only valid when used in combination with the Purchase Success Event.  They will be ignored in all other cases.

Real-World Example
In this week’s real-world example, I will illustrate a non-traditional way to use the Products variable.  In this scenario, a subsidiary of Greco Inc. would like to track the effectiveness of internal promotion banners on their site.  They use these internal promotional banners to promote specific products and for cross-sell.  To do this, they would like to see how many impressions each internal banner received and then how many clicks each received so they can calculate an “Internal Promotion Click-through Rate.”  However, this can be a bit tricky since one page on their site could have multiple internal promotion banners.  While we might normally think about using an eVar to capture the internal promotion codes, since there are multiple codes passed on one page, this becomes problematic.  The best an eVar can do is to concatenate the values into one string, which is not the easiest way to analyze this data.  Therefore, let’s see how they can use the Products variable to accomplish this.

To start, they would need to assign a unique ID# to each version of each internal promotion banner.  When the page loads, they would capture the ID numbers of all banners that appear on the page in the Products variable and set an “Impressions” Success Event.  The tagging might look like this:

s.events=event15;
s.products=”;promo:123,;promo:150,;promo:174″;

Doing this would produce a report that looks like this:

If a visitor clicks on one of the internal promotions (say ID# 174), they would set a “Internal Promotion Clicks” Success Event and capture the clicked internal promotion in the Products variable again with tagging that looks like this:

s.events=event16;
s.products=”;promo:174″;
s.eVar18=”promo:174″;

Then they would create a calculated metric that divides Success Event 16 by Success Event 15 and then they can see a report that looks like this:

As more internal promotions are shown and clicked, this report would receive more and more data so that over time, they could see which internal promotions are clicked the most.  In this situation, I started each product value with the phrase “promo:” since in many cases you will use the Products variable for several things and having this prefix allows you to easily identify which values in the Products report are actual products and which ones are internal promotions.

Finally, it is important to note that if they wanted to see which internal promotion clicks later led to visitors completing Success Events on the site (i.e. purchases, sign-ups, etc…), the Products variable would not be ideal since it is not persistent like a Conversion Variable.  For this reason, you might choose to capture the internal promotion code ID# that is clicked in an eVar (as shown above) so that you can see what internal promotions led to future success.

Have a question about anything related to Omniture SiteCatalyst?  Is there something on your website that you would like to report on, but don’t know how?  Do you have any tips or best practices you want to share?  If so, please leave a comment here or send me an e-mail at insidesitecatalyst@omniture.com and I will do my best to answer it right here on the blog so everyone can learn! (Don’t worry - I won’t use your name or company name!).  If you are on Twitter, you can follow me at http://twitter.com/Omni_man.

Recently, we were working with a client and wanted to introduce them to behavioral targeting for testing. We had already started the initial rounds of tests to targeted paid search traffic but felt there was an opportunity to move the needle a bit more. There are endless things that one can do with behavioral targeting. At a high level though, behavioral targeting allows marketers to target or segment based on pre-defined behaviors of their visitors.

Start Small Enough to Show a Difference

For this particular client, a large retailer, we wanted something relatively easy to do in terms of execution and design so as to deliver a small taste of what is possible. We created a test that would be targeted to those visitors that added something to their cart but failed to complete their purchase, and then made a subsequent visit to a category page, product page, or the home page on the same session or a later session. This segment can be described as “abandoned carts.”

Set Up a Monitoring Campaign First

Now before I we explained to the client what we wanted to do we thought it best to get some data to help explain the goal and our desired outcome. To do this we set up a monitoring campaign to track these cart abandoners. A monitoring campaign is similar to a regular campaign or test but does not deliver content. It is mainly used to identify baselines so as to visualize the gains of sequential A/B or multivariate testing.

For this monitoring campaign to capture the desired data we had to create the segment for cart abandoners. To do this, we profiled or ‘tagged’ those visitors that added something to their cart. We then profiled or ‘tagged’ those visitors that checked out or made a purchase. Profiling is an enormous feature that will certainly be blogged about here soon in much greater detail.

Now, to qualify for the monitoring campaign a visitor had to be a cart-adder but not a purchaser. The visitor would then have visited a category page, product page, or the home page. We did not limit these visits to single session.

The Monitoring Campaign Results

The results of this monitoring campaign were definitely impactful. We found that 30% of visitors finally made a purchase after leaving the funnel. We then had some interesting data to help us understand who we were focusing our test on and what would be our success metric. Our initial goal wasn’t to increase the revenue metrics such as RPV, AOV, or total sales. Our aim was to increase the conversion rate for the 70% of visitors that completely failed to check out despite having added something to their cart and making a subsequent visit back to one of the earlier mentioned areas of the site.

Now that we had the profiles or ‘tags’ in place we were all set. We added an mbox to the top of the center “hero area” to the home page, category page, and the product pages. This mbox was named the same, which allowed for quick and easy set up in Test & Target. The only thing needed now was a creative element to remind visitors that they have something in their cart.

Add Creative

For the creative, we included a handful of different designs to see if there were any differences in the performance. Limiting the test to no cart abandonment messaging vs. one piece of creative limits your analysis quite a bit. If cart abandonment messaging is effective but the creative was poor you would get misleading results in such a test.

The test or campaign was set up as follows:

Recipe A: Default Content (No Cart Abandonment messaging)
Recipe B: Abandonment messaging (design 1)
Recipe C: Abandonment messaging (design 2)
Recipe D: Abandonment messaging (design 3)

With this type of test design we get to understand a couple of things from the results.

1. Does cart abandonment messaging improve the conversion rates of this segment of traffic?
2. Does one creative over another affect things differently?

The results were pretty interesting. We saw a decent amount of conversion lift across Recipes B, C, D. This was great to see, as we were able to confirm that we could positively affect this segment of traffic.

Recipe C though did a lot better then Recipes B and D. The creative was very different which lead us to believe that creative does impact things as well.

Right now this customer is running a champion/challenger test by testing Recipe C against another new variation to see if we can leverage the gains even more. This time we have 10% of traffic going to the control, and 45% of traffic going to each of the cart abandonment messaging. This way we can quantify the impact of the messaging while getting the biggest bang for our buck.

If you want to see a very simplistic mock up of the cart abandonment messaging using the Test and Target platform, I invite you to visit this link.

Once you visit you will see this product page with four products:

If you actually add one of these nice Mac products to your cart and then revisit the product page you will see the tool in action with this messaging:

If you are a Test and Target user and would like to see how I set up this Cart Abandonment test please email me or make a comment on the post. I will then create a login into my Test and Target account. Here you can not only see how I set up this test but the many more examples I plan on posting here.

Understanding Segment Builder
To understand the Segment Builder, you first need to understand the various components that it uses to build a segment.  As shown in the image below, when you enter the Segment Builder, you will see Containers, Events and a Segment Canvas.

First, we will discuss the different containers:

1. Page Views - Dragging a Page View container to the segment canvas will allow you to define which Page Views you would like to include or exclude from the segment.  When evaluating the Page View container, Omniture is, in effect, scanning through each Page View it finds within the specified time frame and deciding whether it should be included or excluded.  Therefore, it may be the case that two different Page Views from the same Visit may or may not be included in the segment.  For example, let’s assume that you are building a segment where you only want pages where the language was Spanish.  It may be the case that a visitor viewed ten pages during their Visit, but only two of those ten were viewed in Spanish.  Using a Page View container, would mean that only these two Page Views would be included in the segment.
2. Visits - Dragging the Visits container to the segment canvas will allow you to define which Visits you would like to include or exclude from the segment.  When evaluating the Visits container, Omniture is, in effect, scanning through each Visit it finds within the specified time frame and deciding whether the entire Visit should be included or excluded.  Therefore, if any of the criteria are met within the Visit, all data from that Visit will be included (or excluded if using the exclude tab) in the segment.  Using the preceding example, if the segment looking for pages viewed in Spanish were built using a Visit container, the entire Visit would be included since at least one of the pages was viewed in Spanish (even though the majority were in English).
3. Visitors - Dragging the Visitors container to the segment canvas will allow you to define which Visitors you would like to include or exclude from the segment.  When evaluating the Visitors container, Omniture is, in effect, scanning through all data it has for each Visitor within the specified time frame and deciding whether at any time the Visitor met the criteria.  If it finds that the Visitor has met the criteria, all Visits and Page Views for that Visitor will be included in the segment.  Continuing the preceding example relating to pages viewed in Spanish, if a Visitor had six site visits within the specified time frame and in one of those Visits viewed at least one page in Spanish, data from all six Visits would be included in the segment.

As you can see, the container you select can have an enormous impact on the data set that is returned in DataWarehouse or ASI.  This is why I cannot stress enough the importance of testing your segments before using them in production.

The next item that can be added to the segment canvas is Events.  The Event containers represent the Success Events you have identified for your website.  If, for example, you have a Visit container and you add  the “Order” Event to it, then you are indicating that only Visits in which an Order took place should be included in the segment.  You can also define the Success Event in more detail, by clicking on its name and adding additional filter criteria.

The most difficult area of Segment Builder to understand is the Segment Canvas.  To begin with, there are two tabs of the canvas, one for Include and one for Exclude.  When determining which data to include in your segment, Omniture first looks to the Include tab and identifies all data that is to be included and then proceeds to eliminate any data matching the criteria of the Exclude tab.  Once you have successfully navigated the include/exclude tab, the next challenging part is to understand how to set conditions within a container.  To do this, you simply click on the underlined container name which brings up a window that allows you to specify the parameters of that container.  Here you can choose almost any data point and specify whether it should be equal to, greater than, contains, etc… and then choose or enter a value.  In addition, using the first drop-down box allows you to choose whether all of these conditions need to be met (AND) or if just one need to be met (OR).  Whatever you choose in this drop-down box will change the criteria in the Filter below.  In the image shown here, the user has chosen to use the OR clause so that a Visit can come from either the United States or the United Kingdom.  One point of warning: The #1 mistake clients make in this step is that they forget to click the “Add” button when adding conditions so be mindful of this!

The final thing to learn in building a segment is the nesting of containers.  Here are some important guidelines you should know:

1. Page View containers can only have nested Events
2. Visit containers can have nested Events and Page View containers
3. Visitor containers can have nested Events, Page Views and Visits
4. If an Event is added to the segment canvas with no Page View/Visit/Visitor container, the Page View container is used by default
5. It is not possible to add multiple Events to a Visit container such that they form an OR clause, but to get around this, you can add multiple Visit containers to the canvas and add one Event to each

The segment Builder will usually tell you when you are attempting to add an invalid combination, but it is good to learn what you can and can’t nest within the segment builder.  Finally, it is important to understand what the Segment Builder treats as AND/OR clauses when building segments.  If you want to add two or more items and have all of the conditions met (AND) then you want to be sure that you nest them above the line of the container like this:

However, if you want to the segment to look for one condition or the other (OR), you would want to make sure that the items are outside of the container line like this:

In the examples shown here, the former would include all Visits where the country was “United States” and an Order took place, but the latter would include all Visits where the country was “United States” or an Order took place.  This is probably the hardest part and I encourage you to tap into Omniture Consulting when building complex segments.

Important Things To Know About Segment Builder
The following are some important things to know about Segment Builder:

1. Segments built using the Page View container will return data the fastest, followed by the Visit container and the Visitor container.  Whenever possible, attempt to use the lowest level container needed to get the data you need to improve processing time.
2. Classifications cannot be used as criteria in DataWarehouse or ASI Segment Builder segments.  For this reason, if you believe that you will want to use a data point as segment criteria, you may want to pass the data directly to a Traffic Variable or Conversion Variable instead of relying upon a classification.  Classifications can be used as segment criteria in Omniture Discover.
3. Segment Builder allows you to segment on how many times a Success Event takes place which is very powerful.  For example, if you set a Success Event each time a visitor conducts an Internal Search, you can build a segment that looks for cases where “Total Internal Searches is greater than 2″ to find repeat searchers.

Tips & Tricks
Since this post is really a subset of the DataWarehouse/ASI posts where we already provided some real-world examples, I am going to provide some practical Segment Builder tips and tricks instead of a real-world example:

• If you want to be sure you are including all data in your segment, you can add a Page View container where IP Address contains “.” since all data collected by SiteCatalyst contains an IP address and all IP addresses contain a “.”

• If you want to build a segment using a wild card such that a SiteCatalyst variable has a value (is not blank), you can use a “greater than !” in your segment definition as shown here.

• Conversely, if you want to build a segment based upon the fact that a SiteCatalyst variable had no value you would use “less than !” in your segment definition as shown here.

Have a question about anything related to Omniture SiteCatalyst?  Is there something on your website that you would like to report on, but don’t know how?  Do you have any tips or best practices you want to share?  If so, please leave a comment here or send me an e-mail at insidesitecatalyst@omniture.com and I will do my best to answer it right here on the blog so everyone can learn! (Don’t worry - I won’t use your name or company name!).  If you are on Twitter, you can follow me at http://twitter.com/Omni_man.

What is ASI?
ASI is a component of Omniture DataWarehouse that allows you to create a data set similar to that of a report suite for a specific segment of users.  Whereas DataWarehouse allows you to build a segment and run a report for that segment, ASI takes this a step further by allowing you to see all SiteCatalyst reports for a segment.  While this may not seem like a big distinction, it is.  ASI allows you to see all Omniture SiteCatalyst out-of-the box reports and the many reports you create by setting Success Events, Traffic Variables and Conversion Variables for any segment of your users that you desire.  In addition, ASI (or Omniture Discover) is the preferred way to view pathing reports for segments of users.

How Does ASI Work?
So how does ASI actually work?  When you create an ASI slot, you give it a name, select or create a segment and then specify whether the ASI should be for a fixed time frame or daily recurring.  If the ASI slot is created for a fixed time frame, Omniture will cycle through all data in that time frame, looking for data that matches the specified segment and include data accordingly.  If you choose a Daily Recurring ASI slot, Omniture will begin with the specified start date and do the same steps as just described, but will continue to add one day of data every 24 hours.

(Click here if you would like to see a Flash demo of the creation of a sample ASI slot)

Once you have submitted the ASI for processing, Omniture will begin processing the data that matches the time frame and segment definition you have specified.  Sometimes, I think of ASI as a time machine that goes back in time and recreates all of the web behavior after the fact for a specific group of people of your choosing.  When ASI has made it through the entire date range, you will be notified via e-mail, though you can use the ASI slot while it is in progress if the time frame is large (one trick I use is to view the Page Views report in the ASI slot which will show you which dates it has processed).  Once the ASI slot is ready, you can use it just as you would any other report suite that appears in your list of SiteCatalyst report suites in the top-left drop-down box.

Important Things to Know About ASI
The following are some important things to know about ASI:

1. You can re-use ASI slots when you are done with them.  Let’s say you need to see a month worth of data for a specific segment, but after you have done your analysis, have no use for it?  Simply purge the ASI slot and after an hour or so it will be ready for re-use.
2. ASI slots can be used to download data to the ExcelClient.  Want to have a data set that shows you all SiteCatalyst reports for 1st time visitors?  Google visitors?  Campaign visitors?  All of this can be easily accomplished via ASI.
3. Any segments that you build in DataWarehouse are available for use in ASI (and Omniture Discover for that matter) since segments are tied to an Omniture user ID.
4. You can manage settings of ASI slots using the Admin Console just as you would any other report suite.  To do this, select “All” in the Report Suites Manager area of the Admin Console.  ASI slots will have report suite ID’s beginning with a “cyg.”
5. ASI’s can be used in A/B comparison reports just like any other SiteCatalyst report suite.

Real-World Example
In this installment of our real-world example, we will be working with Greco Inc.’s banking subsidiary.  In this scenario, the website design group is investigating alternative site designs based upon whether the site visitor is logged into the site or not.  Their hypothesis is that those who are signed in (customers) may be looking for different types of content than new visitors (prospects) such that a one-size-fits-all approach is not justified.  In order to test this theory, the design team worked with the web analytics team to create two distinct ASI slots.  One to contain only data for visitors who were logged in and the other being the converse, containing only data for those who were not logged in when data was collected.

Once these ASI segments had been created and the ASI slots were built, the first thing the web design team wanted to see was what search terms each user type had searched upon.  To see this, they simply open the search term report, switch to A/B comparison mode and select the two relevant ASI slots:

Next, the web design team wants to compare the pathing behaviors of both segments.  The first report shown here provides pathing information on those who had logged in…

…while the following shows the paths for those who had not logged in.

As you can see from these examples, ASI can be extremely helpful in comparing segments and extending your web analysis.

Have a question about anything related to Omniture SiteCatalyst?  Is there something on your website that you would like to report on, but don’t know how?  Do you have any tips or best practices you want to share?  If so, please leave a comment here or send me an e-mail at insidesitecatalyst@omniture.com and I will do my best to answer it right here on the blog so everyone can learn! (Don’t worry - I won’t use your name or company name!).  If you are on Twitter, you can follow me at http://twitter.com/Omni_man.

When you wrote your book ‘Search Analytics - A Guide to Analyzing and Optimizing Website Search Engines,’ what was the most important lesson you hoped website owners would learn?

I wanted to create awareness. I wanted to get people to realize that “site search analytics” (SSA) provides significant insights about the users of a website and their needs.

Before I wrote the book, I used to ask everyone if they did SSA. I have observed that not many website owners were doing it. I have encountered clients who were amazed when I have demonstrated to them how we could turn SSA insights into site improvements and marketing intelligence. So I thought I’d share my metrics and techniques with the analytics community in a book. While writing the book, I have identified a number of organizations who had formal SSA practices. I have also incorporated their experiences into the book as case studies.

Since you wrote the book back in 2006, what has changed in the discipline of search analytics?

There is evidence that more people are doing SSA. At least in Australia, we know this is the case from Bienalto’s Annual Web Analytics Surveys. SSA is consistently ranked in the top analytical techniques used.

I have also noticed that people are asking more questions on the subject in online forums.

Louis Rosenfeld [http://louisrosenfeld.com/presentations/seminars/site_search_analytics/] does SSA workshops and is writing a book.

Also, my book is still selling!

These all tell me there is interest on the subject which in turn would lead to development of useful analytical techniques and improved site search experiences.

You mention in the book that only 37 percent of site searches are successful; is that improving?

I don’t have current research on this. However, I am confident that the search success rate is higher for sites where site search is regarded as an important function of the website and there is a formal SSA practice in place.

On our client projects, we always look at ways to optimize the site search experience and improve the success rate. In addition to fine tuning the indexing rules, addressing content issues, we employ techniques such as facetted search. This allows users to refine the search results, decreasing the exit rates from the results page and increasing the selections.

What best practices do you see driving the improvement, if any?

For organizations that have no or little SSA capability, I believe the best practice is to start with the analytics. I am talking about analytics beyond reporting frequently searched keywords.

For example, through keyword theme analysis on a recent project, we discovered that about half of the searches on a client’s site were for specific brand names. A “long-tail” effect had been created due to the fact that the client carried a large number of brands, prompting a large number of individual keyword searches. The client had previously been unaware of these searches since they did not show up in the top keywords report. In this instance, I am certain that the high concentration of branded keyword searches was caused by navigational issues.

With SSA, the importance of search can be easily established by measuring the proportion of visitors performing searches on the site vs. those navigating directly to their desired pages. Then you can establish what drives visitors to the search. Is it because the site’s navigation fails or is it because your visitors simply favor search to browsing to find information? Finally, analysts can examine how well search results perform.

What is the number one thing website owners should do to improve conversion through search?

Relevance.

Improving the relevance of the search results set and the presenting search results in a way that aids the selection of the most relevant search should be the number one focus.

To do this requires a well-planned implementation of the site search tool (getting the search index right) and, once implemented, continuously monitoring and analyzing how it performs and making necessary adjustments.

Web analytics, as it relates to site search, has been primarily viewed as a tool for understanding search behavior. Do you see the intersection between analytics and site search evolving towards a more dynamic model, one where the site index is influenced by website behavior and vice versa?

When I wrote the book, I argued that SSA is a subset of Web Analytics. Both the context of the search (i.e. the circumstances that lead to search) and after-the-search behaviors should be studied.

I believe there are merits of a dynamic model. It makes sense to focus the search results to the context of the visit. For example, if someone is browsing through the customer support area of a website and conducts a search, it would make sense to display support-related content first in the search results. However, such practices require some caution because the user might have accidentally arrived from Google in a part of the site which is totally irrelevant to what he or she is looking for. Or a site’s navigation may take them somewhere they were not intending to go. Then search becomes an exit path instead of a path to relevant content. I would experiment with dynamic models that apply less rigid rules.

Another area that has merits is using search keywords as part of content targeting to display dynamic content.

What will be the biggest benefits of this evolution to both Web site owners and users?

The biggest benefit will be increased relevance of content and offers. Relevance means engagement, and the engagement leads to conversion and loyalty.

One Site, Many Goals

These are the 3 major objectives I think a campaign site has to accomplish:

1) inform
2) persuade
3) solicit

The problem is that not all visitors progress through each stage, nor do they follow them in any particular order. It’s a bit of a chicken and the egg…and, uh, the cow problem. Luckily, it’s also the perfect scenario to bring your acquisition marketers and site marketers together. Your acquisition channels can help segment your visitors by goal and communicate that back to the site. More on that later through real examples from the Obama and McCain campaigns…

Time = Voters

What I find interesting about political campaigns is the urgency that drives execution at a much faster pace than most companies out there can accomplish today. I suspect that the marketing team behind each candidate ranks near the top for agility and speed because there simply is no other option. The primary problem is not whether there has been enough budget and headcount allocated for 2009, it’s about who bid on the right keywords representing the issues voters are searching for today and who can react fastest to the latest event to hit the headlines.

As we’ve all seen over the last few months, advertising is one of the major vehicles driving a campaign’s messaging and branding. It shows us which topics candidates are focusing on and which channels they are spending the most money on. What’s sometimes missed, though, is what that advertising then tells them about us. There are few advertisements that don’t ask you to do something in return, whether it’s donating on a website or joining a mailing list. Below is a comparison of each campaign’s ad on the same day in the same location, along with the corresponding landing page.

The McCain campaign’s ad on Drudge Report:

Ad’s Landing Page:

Obama Campaign’s Ad on Drudge Report:

Ad’s Landing Page:

You’ll notice that there is a widget available to input information about your income so that you can calculate how much you could potentially save under the Obama/Biden tax plan.

Connect the Offsite and Onsite Experience

Could these two experiences be any more different? McCain’s ad is a straightforward call-to-action to donate. It focuses singularly on the “solicit” goal. However, there are key elements of that ad that are lost once the visitor lands on the site. For example, what does it mean to “Invest in Victory”? Should we at least see that headline reinforced on the landing page, if not elaborated on with how individual contributions can help?

It’s also interesting to note that the call-to-action specifically uses $50 as the amount to contribute, but it is neither highlighted nor pre-selected once the form to donate is presented. I would imagine that some market analysis led to the emphasis of $50 in the ad, so why not continue that through the experience by making the process to donate $50 just a little bit simpler.

In contrast, Obama’s ad is targeting voters based on the issue of taxes. It takes the approach of focusing on the “inform” and “persuade” goals. You’ll see that the landing page calculator asks for anonymous information and then returns an estimate of savings. Here’s what the page looks like after filling out the calculator:

Note that the calculator has changed into a form that now asks for my information to keep in touch. We’re now moving into the next stage of solicitation.

Engage the Visitor

From a design, flow, and relevancy perspective, I think this ad to landing page experience is very well-executed. Beyond that though, I think it’s also a clever way to get a lot of information about your visitors in a way that is not explicit. Let’s take a closer look at what we learn as the visitor progresses through the ad and landing page.

After the visitor clicks on the ad:
We can infer that this visitor is interested in learning more about taxes and saving money.

After the visitor fills out the calculator:
At a minimum, we know the visitor’s annual income and filing status. On top of that, we may know how many dependents they have, whether they are over 65, if they have child care expenses, if they are saving for retirement, the outstanding balance of their mortgage, and if they have college expenses. That is A LOT of information when you look at it from a profile targeting and segmentation perspective. Imagine how much more relevant and engaging you can make this visitor’s experience across multiple channels including display, email, and site by having this type of data.

After the visitor fills out the “Stay Informed” form:
We now have the visitor’s email address and zip code to tie all the information together in the database.

To recap, here’s the wealth of information that the Obama campaign has in their database now after a visitor clicks one ad and fills out two short forms:

• email address
• zip code
• issue interest
• income bracket
• filing status
• age*
• # of dependents*
• mortgage balance*
• saving for retirement*

*optional field

Would you send a different email to the voter in a swing state making $40,000 a year vs. the voter in New York making $300,000 a year?

These Principles Apply to More Than Politics

The strategies employed here are not unique to politics. They might just require a shift in how you approach your business and define the goals of your acquisition marketing efforts.

For example, let’s say you’re a retailer that sells HDTVs. Instead of an ad promoting a specific model, maybe you show an ad that provides information on how to select the right HDTV for you. The landing page then asks the visitor about their budget, preferred size, and other relevant parameters. Maybe that then leads them to personalized search results, or you display another short form that enables them to receive alerts when an HDTV goes on sale within their budget.

Think about how you can make each touch point with your visitor more relevant and engaging so you can draw them through a cohesive experience that both leads them to the water and increases your bottom line. In other words, ask not what your visitor can do for you, but what you can do for your visitor. And don’t forget to vote!

% of Visitors Using Site Search

How often do people use the site search on your site? Given that conversion rate from site search are shown to be nearly 3x higher than average site conversion rates, many sites try to encourage more site search use. If your site search is being used by less than 10% of your visitors, you may want to revisit how the site search function is presented on your site.

Action: Do you have a search box on every page in your site in a consistent location? Is the search box visible above the fold? Could the search form be larger? Does the search work if someone decides to use the ‘Enter’ key after inputting a query term and not the form submit button? You might also consider multivariate testing of the search box location to see how it impacts site search usage. The search box should be a prominent part of your site navigation.

Most Frequently Searched Terms

Look at your top searches and then take a look at the actual search results for your top 25 or 30 terms. Are the top results returned helpful? Take a look at how many visitors are dropping off your site from the search results pages, or are clicking to pages 2 and 3 in the result sets. If some of the top terms are ambiguous as to intent, can your visitors refine the results set to narrow down to a specific topic? Are any “best bet” suggestions offered? It pays to optimize results for your top search terms as our experience is that the top one hundred search terms usually comprise upwards of 40% of all the site searches performed.

Action: Fine-tune your site search relevancy controls as necessary to be sure the right content is being presented in the first few results returned, and can relevancy adjustments be automated based on your analytics data? Ensure that new content is being indexed in a timely manner, and consider adding keyword-based “best bet” promotions in results to help guide visitors. For top terms, also consider building a customized results page that can better serve user intent by aggregating results from different site sections. For example, see how Verizon Wireless aggregates different content from around their site for a search of ‘blackberry’.

Number of Unique Search Terms

Aside from the top search terms, you will also see a large number of unique search terms (the “long tail”). How many unique terms are there in your reports, and how can you optimize the site search experience for all of these terms?

Action: Build a means for your visitors to easily navigate your results sets with appropriate search refinement options. Does your search engine just deliver a generic “Next 10 Results” option as the only means to examine hundreds of search results? More users are expecting to easily narrow results sets to a specific website section, topic, date range or content type. If you provide such refinement options already, how often are they being used? Can the visibility of these options be improved in results pages if the usage numbers look low? See how CBS News uses search refinements to allow visitors to narrow results by site section, type of article, show and date range.

Number of Failed Searches (zero results found)

How many failed searches are happening? Are there top failed search terms? Failed searches usually mean that either your visitors are searching for terms that do not exist on your site (highly useful data in and of itself) or that your current site search solution may not be working properly if the keywords actually do exist in your website content. Scrutinize failed searches for terms for which you do not have content — are your customers telling you something? Do you need to freshen your website content with newer jargon? Are these internal search terms valuable for your SEO/SEM efforts? For failed searches for which you do have content, examine why the failure is happening. Is the content on your site fully searchable? Are the failures a result of misspellings only?

Action: Quickly create synonyms in your site search solution to immediately rectify top failed search terms. Also, be sure that your site search engine has been optimized to handle misspellings, and consider modifying your existing site content to include newer terminology. You might also consider a site search redirect option — for example, if you see searches for “FAQ” but your site does not have an FAQ section, you could instead redirect that visitor to your top-level Support page instead of a failed search page. See how BrooksBrothers supports misspellings.

Measuring and acting on your site search analytics data will increase your conversions. I’ve discussed optimizing results based on a few key site search metrics above and, in future posts, I’ll comment on others as well. Your existing site search solution should allow you to act on your site search analytics data to:

• Easily fine-tune and automate the rankings of results
• Build search refinement options for a guided search experience
• Create customized search results pages for key terms
• Support misspellings, and offer spelling suggestions
• Quickly manage synonyms, acronyms and hyponyms
• Allow keyword based “best bets” and url redirects

If your solution does not offer these options, please give Omniture SiteSearch a close look. Hundreds of leading companies use Omniture SiteSearch to optimize their site search experience including the sites referenced above. All comments are welcome.

What is Omniture DataWarehouse?
Omniture DataWarehouse is a repository of SiteCatalyst data that stores information similar to that which you can see in SiteCatalyst.  While many customers think DataWarehouse is only a backup of their data, it is actually much more than that.  To understand DataWarehouse, you need to first understand how it differs from SiteCatalyst.  When you use SiteCatalyst, the reports you get are pre-defined so that they can return quickly in the browser.  For this reason, you may notice that you can only perform a small number of report breakdowns (Correlations and Subrelations) within SiteCatalyst and that you cannot breakdown Traffic Variables and Conversion Variables by each other.  The reason for this is that these more complex queries could take too long to return, resulting in a browser timeout.  SiteCatalyst is fine-tuned to provide you with speedy access to 80% of the reports you should need on a daily basis.  On the other hand, DataWarehouse stores the raw data which enables it to be used for much more complex queries, but the results are not provided in real-time (normally within 24 hours).

Another key difference between SiteCatalyst and DataWarehouse is that DataWarehouse can provide deeper visitor segmentation.  For example, when SiteCatalyst is collecting information about a visitor, it is difficult for it to know that three months ago this same visitor looked at Product XYZ and that two months ago the visitor began an application but did not complete it.  That level of information requires that Omniture sift through rows and rows of website data which is not easy to do in real-time within a browser.  DataWarehouse, however, contains all of this historical data and has a built-in visitor segmentation engine that allows you to create segments that are meaningful to you and extract SiteCatalyst data for that specific segment.  Common DataWarehouse segments might include:

1. All Visitors who have added a product to the shopping cart, but not purchased
2. All Visits where visitors have viewed a product page, but not added anything to the shopping cart
3. All Visits where visitors came from a few select cities and added a few specific products to the Shopping Cart (see below)

The number of different segments you can create is limited only by your imagination.  DataWarehouse provides a Segment Builder that allows you to choose how you want to build the segment and a canvas that allows you to specify what data you want to see for that segment.  You can also choose the timeframe for the data set and whether you want the resulting report to be delivered one time or on a recurring basis.  In a future blog post, I will cover best practices for using Segment Builder to create DataWarehouse segments since this can be an art in of itself!

When Should You Use DataWarehouse
Once my customers have absorbed the preceding information, the next logical question they ask me is when they should use SiteCatalyst and when they should use DataWarehouse.  Here is how I respond:

1. Use DataWarehouse when you need to see data for a subset of your audience.  If you need to see visitors from the US who have started the application process, SiteCatalyst will be more than adequate, but if you need to see 1st time visitors, using a Google Chrome browser who have started an application at some point, and began their visit on a specific marketing landing page, all for the last 3 months, I would recommend using DataWarehouse (or Omniture Discover).
2. Use DataWarehouse if you need to see Traffic Variables and Conversion Variables broken down by each other and you do not have access to Omniture Discover.  If these breakdowns are going to be needed on a recurring basis, I work with my clients to capture the necessary data into both Traffic and Conversion Variables to avoid having to rely too much on DataWarehouse, but many times this is not feasible and DataWarehouse can really save the day.
3. Use DataWarehouse if you need to go beyond the two levels of report breakdowns offered in SiteCatalyst.  DateWarehouse allows you to create an unlimited number of breakdowns.
4. Use Data Warehouse if you have a piece of data that has more than 500,000 unique values per month.  While this doesn’t happen too often, there are cases where Omniture customers need to pass a user ID or some other unique values to a variable which exceeds the recommended limit of SiteCatalyst.  In these cases, the variable in SiteCatalyst is not useful since it shows a “Uniques Exceeded” value, but all of the data is stored correctly in DataWarehouse where you can build the appropriate segments and extract a list of the relevant unique values as needed.

Important Things to Know About DataWarehouse
The following are some important things to know about DataWarehouse:

1. The more complex the DataWarehouse segment and the larger the time frame, the longer it can take for a report to be returned.
2. The Unique Visitors metric provided in a DataWarehouse report is relative to either the overall time frame or specified report granularity (Day/Week/Month if selected) in DataWarehouse reports.
3. DataWarehouse segments have an “exclude” function that allows you to eliminate data that you don’t want to be included in your segment query.
4. Building too granular of a segment can often times return no data.  My advice is to run a test report for one day of data to be sure you have your segment correct before attempting to run it for months worth of data (learned the hard way from someone who waited a few days only to receive no data due to my own user error!).
5. DataWarehouse can be used to load historical data into Omniture Discover.
6. DataWarehouse is used by many Omniture Genesis integrations.
7. Most of the things you can do in Data Warehouse can be done in Omniture Discover in real-time.

Real-World Example
In this installment of our real-world example, we will focus on the CoolFlowers subsidiary of Greco Inc.  In this scenario, the CMO of CoolFlowers is looking to test a new re-marketing campaign and would like to identify all of its customers who have purchased flowers in the past three months but submitted fewer than three total orders and are from the New York City metro area.  CoolFlowers captures an encrypted customer ID into a Traffic Variable (sProp) on each page after a customer logs into the site and since SiteCatalyst captures orders and geographic location, this can be accomplished through a DataWarehouse request.  However, just to make things a bit more complicated, the CMO would also like to see what Products, if any, the visitors matching this segment have looked at online, how many times and what city they are from within the New York City metro area.

To accomplish all of this, the SiteCatalyst power user (you), would build a segment as described above and then add the necessary data to the reporting canvas which shows what the output file will look like:

Once you have your segment built correctly and have added the correct data on the canvas, you can schedule the report and it will be delivered to the specified e-mail address or FTP site.

Have a question about anything related to Omniture SiteCatalyst?  Is there something on your website that you would like to report on, but don’t know how?  Do you have any tips or best practices you want to share?  If so, please leave a comment here or send me an e-mail at insidesitecatalyst@omniture.com and I will do my best to answer it right here on the blog so everyone can learn! (Don’t worry - I won’t use your name or company name!).  If you are on Twitter, you can follow me at http://twitter.com/Omni_man.