First, you want a Social Media Strategy that centers on the right channels, besides that, you want to make sure that your content is being communicated in a way in which it goes viral in the community and beyond. So, what do we do and what do we need to keep in mind?

A Social Media Campaign should mirror your Social Media Strategy in that you should make it clear, from the start, what your want to achieve. Set your goals! And even though it would be nice to forget about it for one minute, you also need to set a budget, one that is proportional to the goals you defined beforehand. What do you want to achieve and what is it worth to you? Creative brainstorming has the tendency to quickly lose sight of the relationship between cost and benefit, which has repeatedly given us Social Media campaigns that are good, but cost much more than they could ever hope to recoup. I want to stress this point, as there are still way too many large corporations that meddle in the social web without having a real goal or any strategy whatsoever, by throwing large budgets at it without giving a moments though about the returns those campaigns generate. “Everyone's doing it, there surely must be a good reason...”.

Once you have decided on the goals for your campaign and allocated a budget, you can start to brainstorm. There are numerous brainstorming-methods, but the one thing they all have in common, is a set time frame. This is to keep you from losing yourself in your ideas and creativity. It is more effective to set a short time frame and then repeat the process the next day. Use the first brainstorming session to come up with ideas and put them on paper, then, once you have taken a break from it, question those ideas again and put them in concrete terms. The more people you bring to the table, the more opinions you will have. This means that you will never have everyone see eye to eye, which makes it important to make a decision for one idea at some point in time and not mix together multiple ideas because you are trying to reach a compromise everyone can get behind.

Below are four questions you should ask yourself to see if you idea for a campaign can work:

• Is the campaign social?
  Is the content, the benefit or the message of your campaign so relevant, that the community considers it "share-worthy"? How are the users being motivated to share this content with their friends? How can the actual campaign be shared?



• Can you link it to other kinds of media?
  How can you integrate your campaign into other channels and media outlets, so that you enable a cross-channel communication, to strengthen you campaign? By integrating it into your website and customer-newsletter, for example.



• What about scalability?
  How can you communicate your campaign outside of an existing community or your customer-base, in order to increase its reach? By starting a simultaneous Facebook Ad Campaign, for example.



• Is the campaign viable?
  Do you have the necessary technical requirements in place to realize such a campaign? Are there any legal or internal regulations that you need to be mindful of? For example, are you in accordance with the Facebook Promotion Guidelines? Are there possible risks involved with the realization of the campaign and what are ways to reduce them beforehand?

When a searchengine makes a change to their ranking-, analysis- or crawling-algorithm, especially those that apply to a large percentage of the total volume of all queries, you sometimes want to know: “How exactly do they know what is really relevant?”. Well, the short-answer is: “They don't know!”

For the most part, they have a lot of great indicators, but analysing such system-immanent data does not make you omniscient. I will try to make this more accessible through an example and then, I will even try to illustrate how you can use one of the evaluations that the searchengines use to test your own site's relevancy.

First off, let's take a look at the different methods of evaluation, which can be roughly divided into two groups:

• Group 1 is comprised of all the methods that gather and evaluate system-immanent metrics like clickstream-data, CTRs, retention times, semantic analysis, on- and off-site linkstructures etc.


• Group 2 is made up of all the methods that gather and evaluate data that was collected externally (ie. eye-tracking, online-panels, user-testing or focus-groups).

I want you to gain a better understanding of the user-testing side of things. Some of you might think, “Doesn't scale well!” or “Too expensive!”, but I respectfully disagree – and so do searchengines.

A classic example for relevance, where you will only get limited results from things like CTR, retention time, pageload-times etc., are special overlays in the SERPs (Search Engine Results Pages). When we take a look at a searchquery for the stock values for MSFT, GOOG or YHOO, all searchengines will show the most important key-values right on the page. This means that the user will not interact with the overlay in any way that can be measured automatically.


Finance and weather overlays in the SERPs 2012 (Google, Bing, Yahoo!)

“Lies, damned lies, and statistics”

In my time as Search Product Manager, it was not unusual that, out of the blue, an engineer would tell me that the performance of the direct overlays for stock prices, weather, horoscopes, you name it, were not up to par and would therefore be shut down promptly. Personally, I'm sure that something like this would not happen this quickly in an important market as the US, but when an engineer in India is looking at an evaluation of 4.000 CTR-statistics and there are a few features that don't make the CTR-minimums, then I am sure his approach is probably valid, at least from his point of view.

In such a case, where the internal data in a searchengine's system does not yield satisfactory results, you will often move over to user-testing.

I am sure that a lot of readers will be surprised to hear that searchengines also use this method at a relatively high frequency to analyze their own organic SERPs as well as the results pages for News-, Image- or Local-searches. The results that these evaluations yield are taken very seriously and it has already become an industry standard to use these metrics as part of the release-criteria for searchengine-features. The feature won't see the light of day if the it does not get the ratings they were aiming for.

How does such a user-test work?

Simplifying things a little, it can be said that you hire a group of editors/testers and put them in front of a computer, where they will “see” one searchquery as well as the corresponding website that is shown at position X. The tester will then grade the quality of the site in relation to the searchquery, using a simple grading system, comparable to grades in school:


Exemplary mock-up of a testing-environment that a rater would see

Here, a 1 goes above what was expected of the site, down to a 6, which means a site had no relation to the searchquery whatsoever.

To ensure that one rater does not negatively influence the results by making a “wrong” assessment/evaluation, there are usually multiple raters scoring the same query. The final evaluation will then only take those ratings into account, that got identical or near identical scores from testers.

Once the evaluation is done, you are left with a value R, as a description of the relative relevance at the time X. You can use these values and the history these tests entail to figure out trends. Additionally, it is quite common to also test a competitors product in the exact same way during such an evaluation, to allow for a direct comparison.


Schematic description of a test-environment/process

Naturally, there are limits to this method, seeing how it is extremely costly, for one, and there are also certain risks involved. When we take a look at rating long-tail queries, we can assume that, for them, the same holds true as it does for the “Ask the Audience” lifeline on Who-Wants-to-Be-a-Millionaire: nice lifeline for easy questions = head queries, but should be taken with a grain of salt once you get to the $100,000 question = tail queries.

The WWtBaM-contestants are compelled to answer only if they are sure of what they are doing, the same goes for the ranking that the testers use. Looking at the results on the show, though, we notice that – putting it nicely – contestants show difficulties in judging the extent of their own knowledge.

Choosing keyword-sets – or: You can be the best searchengine for brush-manufacturers and still not make a penny.

When we look at topicality, web-search-queries belong to those that are extremely divers. This makes it so much more important to exercise extreme caution when creating a keyword-set for an evaluation of relevance.

It should not come as a surprise, that such benchmarks do not just use random collections of search-queries. When you are looking to get a meaningful result, it becomes necessary to have enough head- as well as tail-queries in your set. Additionally, you want to add specific scores to the search-queries to better make out their values to the tester.

It should be clear to everyone, that a keyword for brush-manufacturers as a subcategory of the local search will receive a much lower score than a keyword for the 'restaurants' subcategory would. This way, they can adjust for the fact that not all areas of search are equally “important”, meaning their ability to be monetarized, their prestigiousness or their relevancy for the product roadmap. Since there is a limit to both time and a budget, these scores help to ensure that investments are made in areas with the largest ROI.

Another challenge in dealing with these query-sets is the act of balancing the scales between the right amount of comparability of data and “up-to-date”-ness. One the one hand, you want to keep the query-set you used for your baseline measurement unchanged, which gives you the ability to get the most exact measure of gains and losses. Having a query-set that is too old, though, will usually have a negative impact on your measurements. If we had a set that, for example, held the query “President of the United States of America”, then the result for the 19th of January 2009 would be completely different from the one for the 20th of January, the day that Barack Obama was sworn into office and became the first African-American President.

So what's in it for me? “Show me the Money!”

Let's get back to the core question that all readers of the SISTRIX blog want answered: how does this help me with my job as SEO, webmaster, product-manager or engineer?

In the most basic set-up, you should grab your mail-man and offer him 50 Euros to sit down with you for 20 minutes. You give him a searchquery, put him in front of your landing page and ask him questions.

This may sound a little “creepy” but hey, “no guts, no glory”, right? I'm sure you know what I mean.

For all of you with a budget largely on the upside of 50 Euros, I would suggest putting some of that money into market research, so that you might get wiser as to the question of relevancy.

I can promise you that, whatever setup you choose, you will learn more on the subject of relevancy than you would perusing the next post on Social Media Ranking Correlation.

Here a tip for those of you not working for a large company with its own testing-lab. Find a market-research firm close to you and get them to put you on their database for tests, focus-groups etc.

I am usually part of such a test at least once a year, and every time I have learned exciting new things. You could consider it advanced training, with the slight twist that, not only is it free, most of the times, you will even get a small monetary compensation – in addition to having good catering and mostly being in air-conditioned rooms.

The dust has somewhat settled on the battlefield that is the Google-SERPs and even though the quality-update only brought changes to about 3 percent of all SERPs, the echo that is heard throughout the SEO-sector is distinctly larger. No wonder – with this update, Google took on those sites and methods of optimization that our craft deals in. We pushed the new data onto our Toolbox today and, looking at our server-load, it becomes apparent that there is a huge interest in analyzing and gaining more information on this subject. Complicating things further is the fact that Google, in the form of Matt Cutts, has, by now, confirmed (here and here) that, not only did they push their new quality-update onto their servers over the past couple of days, they also added a new incarnation of Panda. With this upgraded knowledge, we calculated our VisibilityIndex-lists anew, now only taking into account the changes that happened between Monday (04.23) and Wednesday (04.25) – this sharpens the image and makes the search for the cause more accessible.

While we are on the topic of causes – in their explanatory blogpost, Google already set us on two tracks to be followed: on the one hand, we have the excessive use of onpage-measures (keyword-stuffing) and on the other hand, we have clumsy purchases of links. It can be assumed that Google will likely use more than one indicator to trigger this filter, but as far as I am concerned, I am inclined to see their weighing mechanism more on the offpage- than in the onpage-sector. While you can usually get rid of onpage-mistakes rather quickly, the story changes considerably, when you talk about failings in your linkbuilding. Once a link is set up, it is often quite hard to get rid of it again; and even if you manage to do so, Google is still able to draw interesting conclusions when they look at the timeline of your actions.

When you take a closer look at the backlink-profiles for the domains to the right, you will notice that a surprisingly large number of domains concentrate strongly on “hard” money-keywords in their top-linktexts. One click on the source of these links and you are painted a picture that does not speak for a solid and natural linkbuilding in the home-market for the domain. While Panda dealt more with user signals, it seems that Google went back to relying on “concrete” SEO-criteria for this update. Be it on- or off-page, all of these are things that have been part of a cunning SEO's repertoire for years.

One afterthought on all those voices that speak up after such an update, those who saw this comming from a mile away: Sure, many of the domains in question went against Google's guidelines and Google has every right to take the necessary steps to stop these practices. One thing you should not forget though, is that there are there are numerous sectors where it would be impossible to fairly compete in the SERPs, by only using methods that are within the constraints of Google's guidelines. This makes being gleeful at a time like this uncalled for.

By rolling out a new algorithm-update during the night, after it had been quite on that front for a while, it is likely that Google caused quite a bit of amazement for German SEOs this morning. Additionally, they published a blogposting on the subject to shine some light on the reasoning behind this update, which is to fight webspam – of any kind whatsoever. As far as that is concerned, this update is on par with the Panda-update. While these changes should only be relevant for about 3% of all searchqueries, it can be assumed that extremely competitive searchqueries are disproportionately often the target, which will let the “perceived” rate of change be much higher in the SEO community.

To evaluate the impact on the SERPs, we ran a special analysis of your keyword-base during the morning. We were unable to make out major fluctuations in the SERPs, this time. This would mean that you will need to get used to the way things are now, until the next iteration of the underlying algorithm takes place. Here we have the winning- and losing-domains for last week, based on their VisibilityIndex score:

Winners

#	Domain	Change
1	t3n.de	71%
2	max.de	61%
3	lonelyplanet.de	58%
4	geld.de	56%
5	datingcafe.de	47%
6	nachdenkseiten.de	46%
7	eesom.com	44%
8	geo.de	42%
9	1000kreuzfahrten.de	41%
10	lifeline.de	41%
11	schieb.de	40%
12	internetworld.de	36%
13	wwe.com	36%
14	perlentaucher.de	36%
15	financescout24.de	36%
16	leserservice.de	36%
17	itwissen.info	36%
18	nzz.ch	33%
19	kempinski.com	31%
20	linguee.com	30%

Losers

#	Domain	Change
1	ticcats.de	-99%
2	gutscheinrabatt.eu	-97%
3	lm-boerse.com	-95%
4	tagesgeldkonto.com	-95%
5	prepaid-vergleich-online.de	-93%
6	tagesgeld.org	-93%
7	porno-hub.net	-91%
8	onlinekredite.cc	-87%
9	sinnvollerweise.de	-86%
10	reisen-ito.de	-85%
11	dsl--flatrate.de	-84%
12	besttips.de	-84%
13	finanziert.net	-83%
14	pkv-private-....net	-83%
15	geschenk.com	-83%
16	partnervermittlung-24.net	-83%
17	meine-kontaktlinsen.de	-82%
18	witze.net	-81%
19	gratis-tattoovorlagen.net	-80%
20	golfkurs.com	-80%

At first glance, it seems that Google got exactly the results they were hoping for: sites with high-quality content are returned more often; domains with a multitude of affiliate-links and -content have the tendency to show up less. We can surely expect Google to fine-tune their filters multiple times in the coming weeks and month, their line of approach, however, is set.

The next few days will likely spawn many more articles on possible causes and reasons for the changes in ranking. One thing though that should already be clear is that, in this case, Google is not just betting all their chips on one signal, but weighing and evaluating a host of different references. This makes it harder and harder go gain a ranking-advantage without having the corresponding “substance”, as far as content and user behavior are concerned, backing it up.

Additional posts on this subject (in German):

• Dominik: Google räumt weiter die Serps auf
  
• Mario: Google “doofe Links” Update
  
• Julian: Google spielt großes Update ein
  

Due to the much discussed assumption that Google Social Signals will be valued higher in the future, there is also a higher momentum in discussions about Google Fake Accounts. This led me to take a look at the kind of connections that Google already recognizes and uses in their social search.

Google explains that an online-review by a friend might be more relevant than one by an unknown, which is an assumption that can hardly be denied. Thanks to this, results by established connections to friends are highlighted, which is not always apparent though, as not all of your friends will be actively using Google Plus to publicly review or share content. This means that it is not necessary to have a Google Plus account to influence Social Search as there are a multitude of factors at play:

• Blogs, websites and other content that a users has created or shared


• Pictures that have been shared by your own social connections


• Google Reader subscriptions


• All known web-profiles like Twitter or Flickr


• Content that has been recommended or shared through the use of the +1 button in the SERPs

What's interesting, is the way in which your own social connections are identified. There is, once again, more to it than just Google Plus. All you need to receive personalized results in the social search is a Google Mail account.



By using the contacts in your Google Mail account, Google can identify more connections which, in turn, become part of the Social Search results. This means I will also get the recommendations from the friends of my friends, something that can hardly be controlled anymore but which also opens up the opportunity of influencing what third-parties see. This would mean that if the average number of contacts for a Google Mail were 100, then I could influence my 100 contacts by sharing a link, but I would also influence those 100 contacts that my contacts have, which would mean I would influence 10.000 contacts. On the flip-side, this would also mean that those 10.000 users would be influencing my own Social Search results. This is quite the impressive number, though it is more likely that the actual number is much higher, since Google is also able to identify the contacts on my Twitter account, for example, which are then also added to my social connections. Incidentally, Twitter is once again in a growth spurt in Germany and has reached the 4 million user threshold for the first time.

As far as Google is concerned, they divide you connections into public (e.g. Twitter) and private (e.g. Google Chat) contacts.

If you are interested in why someone is shown in your Social Search results, you can hover your mouse over the name of the person or use the link http://www.google.com/s2/search/social?hl=en to get to the Social connections summary page (the link will only work if you are logged in).

Quite the nice feature, but personally, this also gives me insight into the large amount of factors that need to be kept in mind when setting up a Google Fake Account. The difficulty increases even more if you are planning on setting up a “subject,-” or “expert,-” profile to increase the thematic quality of the profile, as you will need to pick the connections that also have the right topicality. While this is not an impossible feat, it surely will be a challenge and I am already looking forward to the first creative talks and articles on the subject of “setting up social connections”.

The trade in links, in order to manipulate the Google resultspages, is nearly as old as the searchengine itself. While it started out as a reasonably small market segment, with a relatively small effect on the SERPs, it seems that 2007 was the year that Google decided it was a worrisome development. From then on, they have been fighting a relentless war against both linksellers and -buyers. “Google's conscience”, appearing most of the time in form of Matt Cutts threatens, pleads, intimidates – and at the end of the performance, nothing has changed and we get to see the same play again, next year.

A couple of weeks ago, that time was once again here: this year, Google decided to add more of a personal touch and addressed a very large number of Google Webmaster-Tools-Accounts with a short greeting, in which they speak of unnatural links on the webmasters' websites. They did this without getting into specifics (of course), deliberately opting to stay on the side of ominous. They encourage webmaster to get rid of those unnatural links and then ask Google nicely for reconsideration. This measure was accompanied by vague insinuations that links would be evaluated differently in the future, as well as Matt Cutts musings on the subject of overoptimisation of websites. To make sure that everybody is also aware of Google's resolve in that matter, they decided to sacrifice some pawns (appropriately enough in the leadgen-sector, which, in the future, will surely be of no interest to Google).

When I look at the reactions that this years show had within the SEO-sector, it seems that Google managed to be more successful with this approach that in the years prior. Google's deftly scattered remarks are often incorporated without thinking twice about them, links are removed and linksellers are ratted out in the Webmaster-Tools under the guise of self-protection. This, of course, leaves a lasting impression, even on those who, at the moment, do not see a cause for action: maybe there is something to these threats and I might be better off, not buying this link?

Personally, I am looking forward to seeing what play Google will put on next year.

Today, while looking at this week's 'Movers & Shakers” for Germany, I noticed the websites for the car-rental company Hertz. We have the hertz.de domain in the Top-3 winners, while hertz.ch is in the Top-3 on the losing side. This is a great example for how Google deals with geo-targeting for websites.

Movers & Shakers excerpt for CW 16, 2012

First, let's take a look at the VisibilityIndex history for hertz.de: from 2009 on, we can see a rather nice, steady increase in visibility, but also some remarkable pitfalls, especially since August of 2011.

VisibilityIndex history hertz.de (Germany)

Oftentimes, we can explain large breaks in visibility with a failed relaunch (with insufficient 301-redirects) or a Google-penalty. Seeing how we have a continuous pattern of ups and downs on hertz.de, we can be relatively certain that none of these apply here.

These zig-zag patterns in visibility are usually indicative of a domain-wide duplicate-content problem. We can confirm this theory by looking at the following screenshot, which compares the VisibilityIndex history for hertz.de to that of hertz.ch.

Comparison VisibilityIndex history for hertz.de and hertz.ch (Germany)

Whenever we see a drop in the visibility for hertz.de, we have a simultaneous increase in the hertz.ch visibility-score, and vice versa. It becomes apparent that Google is repeatedly having a hard time, allocating these two domains to the correct countries of Germany and Switzerland. For Hertz, this becomes problematic for two reasons. First, they have the problem that, during those weeks, the results for Hertz show up less often in the SERPs. The second problem comes due to the fact that the CTR in Germany should be less for results on hertz.ch than for those on hertz.de. When we take a look at the Toolbox-data for Switzerland, we are treated to the same spectacle on google.ch, with the only difference being the reversed positions of hertz.ch and hertz.de.

A possible explanation for why Google is continuously having the problem of assigning the correct geographic targets to both domains can be found when looking at the backlinkprofile for hertz.ch. There, we can see that the website's links are not predominantly coming from other Swiss sites, but that there are a similar number of strong links coming in from Germany, too.

Overview links hertz.ch

The .de and .ch domains are ccTLDs (Country-Code Top-Level Domain), which means that you are unable to change the geographic target for these domains in the Google Webmaster-Tools (GWT). This is something, that can only be done for generic domains like .com or .eu or for individual sub-domains and directories. Google notes that the country-specific domain-endings are one of the signals they use, when figuring out the geographic orientation for the website, while others are the server-location (IP-address for the server) and additional indicators. The above example for Hertz shows that there is still a potential for problems, when it comes to geo-targeting. This goes hand in hand with Google's statement that “Geotargeting isn’t an exact science, so it's important to consider users who land on the "wrong" version of your site.” It is also important to note, that Google made the announcement that they are not using any code-level language information such as lang attributes, or Meta-Tags such as geo.position or distribution.

Going back to the Hertz case, one way to fix the problem is to use the rel=”alternate” hreflang=”x” (not to be confused with the lang-attribute) annotations in the <head> HTML portion of the website, where you can use the hreflang attribute across domains. This means that you can use the hreflang attribute de-DE (German content for German users) and de-CH (German content for Swiss users), for example. This enables the website-operator to point to the correct version of the website and for Google it becomes easier to understand the relationship between the respective domains. Detailed information on how to use hreflang can be found on Google's Webmaster-Tools-Help page.

When you look back at the SMX Munich, you will notice that there were very few presentations that did not, at least, touch on “Social Signals”. This surprised me a bit, seeing that the question of whether social signals are relevant for searchengine optimizations is one that, one year ago, was cause for controversy among SEOs. I remember talking to SEOs that disputed the idea that these signals had any relevancy whatsoever and who were sure that, “SEO can be done without Social Media”. I started getting into Social SEO in 2009 and I am pretty sure of its increasing relevancy, especially now, with Google making its Google Plus network their highest priority. Another factor to keep an eye out for, is the new social media star “Pinterest”, which is not only one of the top traffic-boosters, but also “still” handing out follow links.

A few weeks ago, I started listening to the “Google Story” as an audiobook, which took me back to Google’s beginnings. With everything we do in SEO, we should never forget Google's main idea: ”Return the most relevant results for the users searchquery”. Originally, the web was a place where the user needed to know a website to get to it – or he was able to get there via a link from a website he already knew, that recommended that site. This means links are basically a form of recommendation, which Google uses to calculate the relevancy of a website. There was no other “currency” out there – that is, until Social Networks came along. There, users are recommending content on the web, all day long – this makes it seem natural for Google to incorporate these recommendations into their algorithms, “to give the users the most relevant results for their searchqueries”.

This means that, sooner or later, everyone who has anything to do with searchengineoptimization will not be able to avoid the subject of Social Media – for some keywords more than for others. OK, it is not likely that every subject matter will be discussed on social networks in the future, but many topics are already being discussed and “recommended”, extensively, many times a day, on social networks. There are numerous companies that were only able to increase their visibility and reach because their content was shared through social media. Not every product is being searched for actively, it may just be that we do not yet know they exist. Social networks make it possible to create a demand and then spread it virally. This viral distribution is the high art of creating reach, inspiring discussions and, by doing so, generating links, because others will report on it. This means that SEO and Social Media are moving ever more closely together.

As far as this relationship is concerned, I learned to love a special function in the SISTRIX Toolbox.

At https://tools.sistrix.de/social/changes (only accessible when logged-in), you are shown the “Top Social Signal” winners for the day. This makes it a great compendium of best practice examples for good social media campaigns and viral distribution. One feature that I enjoy immensely is the visibility history, combined with social signals.

At first, you notice the usual suspects, YouTube, Facebook, BBC, Stern..., but you have the ability to sort by networks. It starts to get especially interesting when you start sorting by Google Plus, as you will notice which companies are already successfully using Google's new service. At the moment, its quite hard to find Google Plus best cases, as they are still rare and hard to find.

I decided to take Pizza Hut as an example, as I asked myself where they got all their Google Plus Ones from? When I looked into this matter, Pizza Hut managed to gather a total of 850 Google Plus votes in the SISTRIX “Social Winners”-list. Pizza Hut does have a Google Plus account, but one with no posts on it and which is not linked to on www.pizzahut.com.

Daily Social Signals winners (sorted by number of total Google+ votes)

In conclusion, this would mean that the controversial Google Plus button on the Google SERPs is actually being used. As far as I am concerned, I was always skeptical of this, because who would give a “+1” to a site that they have not looked at at that time, or who comes back to Google to click on the “+1” button? Now I know that, for Pizza Hut at least, it's 850 users who do.

+1-button on the SERPs

Google can use these votes to identify brands, seeing that it is rather likely that only well known brands, with which users can identify with or which they love, will be able to get these numbers of “+1” votes.

Just in case you are wondering; all those tweets, that Pizza Hut managed to rack up to become one of the Social Winners for that day, are due to the ability to tweet about a successful online-order from Pizza Hut: http://topsy.com/www.pizzahut.com.

Another interesting example is the US TV series “Smash”. As a short intermission, here is a good place to call attention to the possibility of sorting the results by country, which is surely another way to arrive at interesting conclusions about the social winners.

Now back to Smash. If you click on the + behind the domain in the toolbox, you are shown the graphical representation of the history for the social reach. This is also where you will notice that the social reach for Smash increased greatly with the start of its first season.

Chronological history social signals Smash (Source: SISTRIX Toolbox)

This may seem a little like the age old question of what came first, the hen or the egg? I would like to wager the guess that the strong viral distribution is part of why the series is so successful and has such reach. A series without constant success would generate a lot of links at the beginning, but would stagnate over time. The idea, that the social votes mirror the series success, can be confirmed when looking at the numbers for the series on Wikipedia.

More evidence for the series success can be found in the SISTRIX Toolbox at https://tools.sistrix.de/domain/nbc.com/social. The page for the series is one of the top social media URLs on nbc.com.

Now that the SEO Campixx in Berlin and the SMX Munich are over, April will give us a great opportunity to get the next SEO-regular's table in Bonn on its way. The plan is to have a cozy get-together on Thursday, April 26th 2012. As always, the regular's table will start at 7pm MEZ. Everyone interested in SEO is cordially invited to attend.

To sign-up, please use this form. We will send you all the necessary information about the location a few days prior to the event. Please remember to sign-up soon, as there is an attendance limit of 50 people.

Some of you might have noticed that we have had some changes this week in the VisibilityIndex scores for both keywords and websites in the local search. This is due to the way the Toolbox handles the local-hybrid-results (blended places search).

These examples should make it easier to understand the difference between a normal Universal-Search integration and a local-hybrid-result: Here the example for "hybrid-local-results" and here the example for a "Universal-Search-integration". At the moment, we are only able to differentiate between the two types, both visually and in the source-code, through the heading “Places for ...”. When we look at the second example above, “radiologie hamburg” (radiology in Hamburg), we are shown the heading “Places for radiology near Hamburg”. Until a short time ago, Google was also adding a snippet to the hybrid-local-results, identical to the organic results, which made differentiating the two much easier. It seems that this snippet has now disappeared from all hybrid-local-results altogether.

There is another important difference between these two result-types: when it comes to local-hybrid-results, there is a very high chance that Google will show an organic result as the first result in that block, but display it like they would a places-result. Often times there is more than one organic result in these blocks.

How can we tell? Simple, just run the same search on T-Online or AOL. Both are in a search-partnership with Google, but they will always display the results without a Universal-Search integration. This means they will show the pure, unaltered organic results. If you want to read more on this, here and here are two blog-posts I wrote on the matter.

Let me use today's searchresults from both Google and T-Online for “sprachschule münchen” (language schools in Munich), as a fresh example. I put both results in a table to make comparing the results more accessible.


As you can see, the first four results (sprachschule-aktiv-muenchen.de, inlingua-muenchen.de, berlitz.de und tandem-muenchen.de) for the hybrid-local-results are identical to the organic results we get from T-Online and the fifth spot (aslsprachen.de) can also be found in T-Online's Top-10. It seems that Google was able to clearly identify a specific Places-profile for each of these five organic results, which they then combined into one block of hybrid searchresults.

When we look at positions number 6 (sprachdirekt.de) and 7 (sprachschule-cosmopolitan.de), we notice that they are absent from the T-Online Top-10. It is likely, that Google padded the "7-pack" of the hybrid-local-results with additional Places-results. Results 8 (prisma-ev-de), 9 (deutschakademie.de), 10 (desk-sprachkurs.de), 11(dialog-sprachschule.de) and 12 (muenchen.de) can, once again, be found in T-Online's Top-10.

When we include the hybrid-local-results to the mix, we notice that the Google-SERPs have two more results than the SERPs for T-Online. These two results are an exact match to the number of Places-results that Google is presumably adding. Apart from that, both SERPs are identical.

This leaves us with a giant question mark. How should we handle hybrid-local-results in the Toolbox? Seeing how our customers use the data on a day-to-day base and report them to their customers, we want to know what they are saying. This means that today is your day. We want to hear your thoughts on this matter.

We believe that it would be a mistake to completely label these hybrid-local-results as a pure Universal-Search integration, without counting them as an organic ranking. A site like berlitz.de is obviously an organic Top-10 ranking for “sprachschule münchen”, even when Google decides to show this result differently from other organic results. Many SEOs will certainly be at a loss for words when they are asked why the website they are in charge of can not be found in the organic Top-10 ranking for important keywords, if we would not count these results. For that reason we started to incorporate them this week as a test-run.

When we treat the hybrid-local-results as organic results, we will count some results that are not really organic results at all. Such results are shown only because Google needs to fill up a "7-pack" or because they believe that they can guess the city-of-origin for a non-city-specific searchquery. These results are not something we are strictly happy with and they are also the reason for the changes in VisibilityIndex score for the Movers & Shakers that could be observed this week.

It is not really possible to clearly attribute these results to one category or another because, just as the name already states, these are hybrid results, with the addition of there also being some exceptions, too. It seems that we can make out the trend that hybrid-local-results have become important, even though, at the moment , Google is obviously still experimenting a lot with the integration and how to display the local results. Google's so-called Venice-update also seems to be indicative of this.

Our suggestion is that we will count the hybrid-local-results as both organic- as well as Universal-Search-results and clearly label them as such in the Toolbox.

Ultimately, it seems that we now have a third established type besides the organic- and Universal-Search-results, that does not fit into any classification system we have for SEOs and SEO-tools at the present. We are looking forward to your input.