When we first launched m.toyota.com in December of ‘08, we were excited to bring a better experience to the many mobile users already accessing the desktop Toyota website. Despite the growing prevalence of full-HTML mobile internet devices (MIDs), we saw a real value in providing Toyota customers with fast and easy access to contextually relevant content - something a desktop site simply can’t provide, even on an iPhone. The results didn’t prove us wrong. The escalation in traffic from 14,000 to more than 40, 000 unique users per week within six months, all from natural search alone, validated our suspicion that auto consumers often have a pressing need to access their favorite brand websites on the go.

But it was what we saw them doing - viewing the dealer locator - and when we saw them doing it - evenings and weekends - that was really interesting. It was clear to us from their behavior that Toyota’s mobile users were deep into the consideration phase and primarily using the site as a shopping tool. So providing them with the best possible shopping tools was the logical next step but that’s where it gets complicated. More than half of m.toyota.com traffic comes from iPhone and iPod Touch with the rest a mixed bag of less popular 3G devices like Android and Blackberry and 2.5G feature phones like the Motorola Razr. Each are equally important groups but with very different capabilities and successfully serving the needs of both without compromising the user experience for either is a tremendous challenge.

Disparity of devices is actually one of mobile’s biggest challenges and probably the main reason many brands shy away from it altogether. Most aim for the middle with a single site that adequately serves the 2.5G users but presents a rather lackluster experience for 3G. To us however, it was obvious we’d be wasting a golden opportunity if we didn’t enable 3G device owners to use their more sophisticated browsers and multimedia capabilities to enhance their shopping experience. The ability to get more detailed specs, view model details, see video clips - all small but valuable elements that would facilitate the final purchase decision.

Targeting the iPhone and iPod Touch was foregone conclusion - with more than half of m.toyota.com traffic coming from these devices, it was clear that these users were a force to be reckoned with. Prius content was also an obvious choice since the demographics for car and phone match up rather neatly. Both brands epitomize a new way of viewing and using technology in one’s daily life and our newly launched Prius experience for iPhone uses the capabilities of one to showcase the benefits of the other. iPhone users visiting m.toyota.com that venture into the Prius section of the site are now treated to an immersive brand experience that offers 360 degree model views, a colorizer, video clips, feature demos and quick and easy access to dealer info - not to mention links to relevant social spaces. All tools that we hope will render the process of choosing and purchasing a Prius faster and more enjoyable than ever.

With the way the web is evolving, the one-size fits all model is swiftly becoming a thing of the past. Both Toyota and iCrossing are look forward to creating new interactive experiences where your content is relevant to what you want, where you are and how you’re accessing the web. Prius for iPhone is just the first step towards an optimal user experience for all.

Rachel Pasqua is Director of Mobile Marketing for iCrossing.

So which social media sites refer the most?  Here is my list of Top 20 sites, ranked by the number of direct click-throughs they provided in June 2009 to a panel of 100 domains in iCrossing’s proprietary data warehouse.

Domain	% of Referrals
facebook.com	46.70%
blogspot.com	15.35%
wikipedia.org	15.04%
photobucket.com	5.15%
linkedin.com	5.11%
myspace.com	4.94%
yelp.com	4.76%
wordpress.com	1.27%
youtube.com	0.75%
livejournal.com	0.38%
flickr.com	0.14%
orkut.com	0.11%
twitter.com	0.09%
shutterfly.com	0.07%
topix.com	0.07%
bebo.com	0.03%
hi5.com	0.01%
reddit.com	0.01%
revver.com	0.01%
omgili.com	0.01%

Twittter was a surprise here, with its rank coming in much lower than I expected.  We did some testing to make sure that the shortened URLs weren’t masking its traffic and found that the 1% share was accurate.

Once again, direct referrals are only a small part of the total value that social media provides.  But a direct referral is a welcome event for any brand since it represents an informed, motivated visitor.

Stan Pugsley is director of business intelligence for iCrossing.

“Faceted search enables users to navigate multi-dimensional information space by combining text search with a progressive narrowing of choices in each dimension” -Association for Computing Machinery’s Special Interest Group on Information Retrieval.

There are various ways to achieve this functionality, one such easy and effective way is using Apache Solr. Solr an open source enterprise search server based on the Lucene Java search library, with XML/HTTP APIs, caching, replication, and a web administration interface.

(Click image to enlarge)

At a high level you can implement faceted search for your website or application by following these simple steps:

1. Download and setup Solr.
2. Configure and setup schema, reflecting your business model.
3. Create data set in the form of XML according to the schema defined.
4. Post the data set to Solr server.
5. Configure queries to get the results as facets.
6. Configure user interface to display the results from Solr server.

That’s it! Follow those steps and your site will have cool-looking faceted search. Well, for a production system you need to run Solr on a proper application server like Tomcat, JBoss, Websphere, Weblogic etc. If the expected traffic is lot, Solr instances have to be scaled accordingly. Data from Solr servers can be exposed as Web services if required, and Lucene indexes can be updated  with additional data by HTTP request or by directly posting data to the server. I have provided a high-level view of what Solr can do, and more information can be found on the Solr site.

Satish Tallapaka is Director of Technology for iCrossing.

For search advertisers, this means that you will soon be able to extend your existing Adwords buy into various mobile applications for the iPhone, and others. Other benefits to advertisers include extended reach on mobile networks, targeting traffic from specific applications and locations, and of course, targeting by specific keywords and categories.

Here’s a case study on Urbanspoon.

Rob Garner is Strategy Director for iCrossing.

Looking at Google Brand CPCs for 22 of our largest clients, we saw that average CPC rose 6%, from $0.29 to $0.31. Defying our predictions, click-through rates actually rose very slightly to 17.21%, up from 17.1% the week before. When we look at Broad and Exact performance separately, we see that Broad CPCs fell almost 6% while Exact CPCs rose 6%. Correspondingly, CTR for Broad was up very slightly while CTR for Exact fell by about the same amount. For our date range we used the first four days following the policy change vs. the first four days of the previous week.

So what does this tell us? One of the things we have seen this week is that a lot of advertisers who COULD be legitimately bidding and using trademarked terms in ad copy (hellooooo, eBay?) are not for some reason. We have seen some folks, like Amazon, starting to use trademarked terms in ad copy but not nearly as many as we would have guessed. Looking at SERPs for major brands this week, it is still very common to see only a single ad for the brand owner with no other advertisers present at all.

Is this all about the advertisers or does this have anything to do with Google? It’s certainly possible that Google is rolling out the change slowly so that they can monitor and tweak the system. It’s also possible that Google is patiently waiting for industry folks like us to do our analysis, declare that nothing has happened, and then not think about it again for several months while they slowly change the serving. (Do I sound paranoid? Okay, maybe a bit.)

Another theory is that Google is primarily making this change so that major resellers (think Best Buy, Walmart) can clearly advertise that they sell Sony TVs as opposed to, say, eBay being able to yammer on about how they resell used items from Sony, LEGO, DKNY, Kenmore (and literally every other brand you can think of). I think the latter scenario is less appealing as a user experience so I’m hoping that this is the case. I think we all remember (and would prefer to forget) the days when eBay used the dictionary as their keyword list.

And at the end of the day, it might simply boil down to the fact that a few days isn’t long enough to uncover anything conclusive, even across so many clients and verticals. So for now we’ll call it a draw, wait and revisit this test in another month or so.

Reid Spice is Director of Search Media Strategy for iCrossing.

But Google’s dominance is not monolithic.  Around the world, and among various user communities, Google’s influence varies widely.

Geographical Variations:

According to one survey of 48 countries, Google’s market share ranged from almost non-existent to almost complete monopoly.  South Korea had the lowest rate of Google use, with only 3% market share.  Lithuania, on the other hand, topped the charts with a 98% market share for Google.  For the countries represented, Google had a median share of 81% of searches.

User Community Variations:

Even within the U.S. market, there are wide variations in Google use.  Using proprietary data from iCrossing’s data warehouse, I looked at the market share across a sample set of 200 million search visits to dozens of domains over the last two years. I found that Google had an overall market share of 72%.  During the month of May 2009, the overall market share had increased to 74%. These numbers align well with the Hitwise estimates.

Among domains, though, the market share varied widely, ranging from 55% to 87%, with a median of 72%.

Financial Services: For visitors to financial services domains, Google averaged 62% marketshare. Yahoo! showed significant strength, most likely due to the popularity of Yahoo! Finance.

Health Care: 65% of search visitors came from Google, also lower than the market average

High Tech: 85% of search visitors came from Google, most likely due to Google’s strength in technical search.

Travel/Hospitality: 78% share from Google.

Retail: No difference from the market, with 73% share.

This is a small sample size, but for the search engines it does show an opportunity to segment users and exploit their interests in activities that are naturally associated with search (such as tracking financial markets with Yahoo Finance).

The implication for brands and marketers is less obvious. A visitor is a visitor, no matter what the engine. At least in the U.S., Google remains the clear starting point for any site seeking to improve SEO or buy keywords.  But Google should never be the stopping point.  Other engines may deliver lower volume, but better quality visitors or better return on ad spend.  Financial services and health care organizations should especially look beyond Google - and treat the total U.S. market share numbers with healthy skepticism.

Stan Pugsley is Senior Director of Business Intelligence for iCrossing.

This got me thinking. I wonder how many people are using Bing’s “Related Searches” feature to refine their original search query and find our clients’ websites? It’s relatively easy to find out, all you need to know is what to look for. When you click on a “Related Search” in Bing after your initial search, a unique parameter is appended to the URL (=R5FD1, for example), which is very easy to locate when looking at a list of referring URLs in your analytics application. Interest2action, iCrossing’s proprietary SEO application, has a top referring URL report so this information was very easy to track down.

I analyzed data for a seven-day time period, across a variety of clients in different verticals. Here are the results:

Client 1: 568 of a total 15,290 referrals, or 3.7% of all Bing referring URLs, came from “Related Searches” or included R5FD in the string.
Client 2: 429 of a total 10,943 referrals, or 3.9% of all Bing referring URLs, came from “Related Searches” or included R5FD in the string.
Client 3: 153 of a total 5,539 referrals, or 2.8% of all Bing referring URLs, came from “Related Searches” or included R5FD in the string.
Client 4: 63 of a total 1,964 referrals, or 3.2% of all Bing referring URLs, came from “Related Searches” or included R5FD in the string.

The “Searches related to” feature in Google has been available for quite some time now, but I don’t think I have ever used it, not even once (for searching purposes). I pulled down referring URLs from Google using the same criteria as above (seven-day time frame, same clients) and the average was less than .4%. While Bing may never catch Google when it comes to search engine marketshare, they have the upper hand when it comes to related searches.

Collin Cornwell is Vice President of Natural Search for iCrossing

“So, Naked Pizza (ahem) erects a Twitter billboard; so what? What’s the big deal?” Where to start… Where else can you establish a direct, sans middle-man, one-to-one relationship with your opt-in customers, in real time, for free (at least for now)? How many other sites garner this level of ingenuity, dedication, and love from techs to stars? There must be something to this Twitter thing, right?

Have you tried Twitter? Do you not get it, but haven’t asked for help from those who do? What will you say to a client, or industry peer, or the media who asks you about it, if you use it, what its value is, and what you’re doing with it?

If you’re one of those folks that still thinks Twitter is just about people telling each other what they ate for lunch, wake up or you’re going to be left behind. I urge you to take a deeper look at it with a fresh perspective. Yes, you’ll still find disposable discussions, meaningless quips, and and pointless posts, but there are also real discussions being held 140 characters at a time; it’s where the conversation is happening. Maybe not tomorrow, but definitely today.

Does anyone else see the shape of the future in the real-time search platforms being built on top of Twitter’s real-time conversation engine? Does anyone else see URL shortening sites like bit.ly (et. al.) rising in number and popularity all of a sudden due to Twitter? Does anyone else see the sheer number of new sites and applications springing up all over the place due to Twitter’s robust and flexible API?

I talk a lot about Twitter, but really, it’s just a name… a label for a rapidly growing conversation, direct-engagement, real-time, instant search, cross-channel communication, data storage, unstoppable monster (in a good way).

Update: In a recent article, Ad Age looks back on the experiment, and declares it a success.

Matt Schultz (@byteengine) is VP, Technology for iCrossing.

+  MSNBot is still the user-agent that is spidering and collecting web data. There should be no impact on sites using any form of user-agent detection.

+  Based on several tests using a variety of non-braded keywords, visibility is fairly consistent with what we were seeing historically on Live Search.

+  Bing is storing your search history “in order to provide relevant results while improving the quality.”  According to its help page, your search history is stored for the duration of your current browser session.  This enhancement will further dilute position data, similar to what has happened on Google over the last 18 months.

+  Many of the refinements Microsoft made in switching from Live Search to Bing are to enhance the decision-making process of searchers.  Among these refinements is the addition of categories that display for many high-level search queries.  These categories offer popular refinements of search queries, and often appear in the search results.

+  Bing intends to deliver more data in the search results page.  From customer service phone numbers and vertical-specific content like Microsoft’s Farecast for airfares, to full Mayo Clinic articles for medicine and enhanced local pages with reviews and travel booking information – expect users to spend more time on Bing and less time on result sites for those topics.

+  A “rich-get-richer” scenario emerges as more of the top listings from related searches are pushed to the front page. This is a smart refinement, considering that multiple research providers have proven that the vast majority of natural clicks occur in the top 5 results. This renders a 2nd or 3rd page Bing ranking much less valuable in terms of its ability to be seen or drive traffic, making things much more difficult from an SEO perspective.

+  Overall, several of the enhancements are borrowed from the Google Universal model.  Results from various proprietary vertical engines are pushed into the forefront of Web search, and while many of the standard optimization principles for images, maps, and video apply, the increased placement of these assets on mainstream web search warrants greater SEO scrutiny of Bing’s ranking approach in these vertical areas.

+  Next to any result, Bing offers a preview of text on the page and displays popular links as well.  Users could theoretically see one page in the search results and click a Bing preview link to another, deeper page on your site. Click-through rates could also decrease on pages that don’t have copy relevant to the user’s query or links to sub-pages the user may be interested in.

Overall, Bing represents a significant enhancement in functionality and direction over Live Search, and should be a more formidable competitor in the search market, especially with Microsoft’s backing.  Early sentiment is very positive for Bing among knowledgeable searchers and search marketing professionals, but it remains to be seen if mainstream searchers will adopt enough volume to move positive market share. iCrossing believes that Microsoft has developed as a much stronger player than Google in terms of relevancy and the quality of the results, though it will likely take search market share at the expense of Yahoo!, Ask and AOL, rather than Google.

Collin Cornwell is Vice President of Natural Search for iCrossing

In all, competition between Y!, MSN, and Google benefits consumers and advertisers as the result of this competition should be a better search experience with stronger consumer engagement and utility. Incorporating advertising programs into this experience will always be a focus due to its revenue potential, however, leading with the best user experience first, and then building advertising solutions into that experience has proven to be the model for success.

The following are a few key, initial points of consideration in regards to Bing. Keep in mind it’s early and refinements are likely but given what is currently available in the market, here are the key considerations:

1.   Live Search is dead. Microsoft is sending all search traffic to Bing – even when you type in www.live.com in the URL bar.  Whatever impact Bing is going to have in the long run, it’s the search engine getting all the traffic that Live Search got just last week.

2. Bing almost always shows highly visible links to competitors on branded searches. Almost all branded searches have a “similar to this” section on the top right, where it looks like the brand’s top competitors are listed this is likely to decrease CTR on branded terms and increase comparison shopping by users.  Furthermore, it’s clear that it is focusing on the most significant competitors – so the consequence is not going to be the same for different brands. For larger, more established brands, they would be more likely to show on competitor branded searches than lesser known brands.  So lesser known brands will have a reduction in branded click traffic because of an increase in diverted traffic to competitors, but will not have a corresponding increase in traffic from its competitors.  In other words, smaller brands get hurt and bigger brands get helped by this system.

3. Bing often limits or downgrades the visibility of sponsored listings. It looks like Bing is still in a very experimental stage, so the SERP varies significantly depending on the search, but it appears that in many cases the sponsored listings have more limited visibility than they did on live search. On several branded searches we performed, the paid listing was not showing on the top of the page, but instead far down the page on the right hand side – well below the fold.  Further, the paid listing was far less visible than the “similar to this” links. This is area where we believe we will see refinement in the user experience and SERP page construction. We believe expect the outcome will be that Bing will balance consumer engagement and utility with appropriately prominent placement of paid search listings that are relevant and useful to consumers.

4. Bing encourages search refinement. This change, long touted as one of the primary purposes of the search engine, looks like it will have less of an impact on paid search than the other changes above, but it does look like users will be more likely to refine their searches on Bing, which should increase searches on more specific terms and give increased importance to the management of tail keywords. Further, the way in which Bing handles paid search listings over the course of users’ search refinement may prove to be a key performance driver or limiter.

5. Personalization. Bing uses user IP address and associated user behavior as an element in deciding which ads to display. Although the intention here is to provide the “right” ads the “right” users, the application and utility of this may prove to be very difficult for Bing and thus both users and advertisers. The concern is that users who behaved one way in order to accomplish one search may behave very differently when trying to accomplish an alternate search. Personalization often struggles to consider and adapt to user intent. Consider how search behavior may change when a user is looking for driving directions versus a user intending to purchase a product online.

Clearly we’ll know more when we start seeing performance data, but from here it looks like there could be some substantial changes to sponsored listing performance.

Chris Wallace is Senior Vice President of Digital Media Services at iCrossing.