Google had an attractive model when it when it first went public. In fact, it was a cash cow out of the gate. Comparatively, Facebook’s revenue prospects right before going public are good but not as clear. As an online marketer who has bought advertising from both Google and Facebook since they both began selling it, I can affirmatively say that what Google offers is phenomenal and what Facebook offers is ho-hum, so-so, really not that great. Google continues to bring in new buyers at efficiency rates as good as and often better than that of the top tiers of the other direct response channels. Facebook, on the other hand, has not impressed many marketers on the advertising side from a performance standpoint.

Apart from the revenue Facebook makes from selling its credits, the bulk will come from advertising. Perhaps there will be a subscription model of some sort down the road, or fees for somehow monetizing their trove of data, but for now, advertising is expected to be the largest piece of Facebook’s revenue pie.

As mentioned above, there have been performance-related issues with Facebook advertising. We are hopeful that Facebook will crack this nut, and if so, marketers, agencies, and of course, all of the Facebook investors, will win big. Here is a brief rundown of why Facebook will be revenue challenged on the advertising side based on their offering as it stands right now:

1. Facebook CPC ads do not produce scale for advertisers due to low conversion rates, high CPCs, and excessive time spent needed to refresh ads that fatigue quickly. Advertisers are attracted to the targeting Facebook offers, which are enabled by the mass of knowledge they have on their users. However, with no clear cut intent to purchase on behalf of the Facebook users when on Facebook, Facebook ads in their current iteration will continue to perform mediocrely on a direct response basis.
2. Mobile advertising, in general, has so far produced terrible performance results regardless of media source or provider. Clicks from mobile display ad networks are cheap and plentiful, but they don’t convert. Mobile paid search ads on Smartphone’s result in conversion rates that are a fraction of what the same ads yield on tablets and desktops. Facebook’s growth is reliant on the mobile platform. These issues combined with #1 above create big challenges for Facebook.
3. Facebook stores do not yield sales. There was a rush by brands to put these pages up under the belief that their customers were spending so much time on Facebook that they would segue from the brand page to the store tab to complete purchases. As it turned out, nobody did. Note that Facebook never took a cut of the sales coming from these pages (there was speculation early on that they might), so the lack of performance of Facebook stores is not in any way a reflection on Facebook’s advertising offering. The point of mentioning this is to reiterate the lack of purchase intent by users when on Facebook. Knowing that puts the onus on Facebook to create compelling ads that convert.
4. In a word: where is the agency support? Pairing up with agencies is a one-to-many approach that Facebook needs to embrace. Google’s agency support is plentiful, and as a result, has enabled paid search and Google Display Network campaigns to become quite sophisticated and considerably larger than they might have been without it. Arming agencies with sales tools lets them rep the media provider (for free). There is no way for a media provider to lose here. Facebook, are you listening?

A revenue stream we suspect that Facebook will articulate soon is charging brands to increase the visibility of their posts. With the advent of Timeline, we have seen a wide scale drop off in visibility of brands’ Facebook pages, and most of the brands we’ve spoken to will do what they have to get their visibility back up. We suspect this drop off is an advertising play, for sure.

Generally speaking, Google’s “Penguin” release is a change in their algorithm aimed at penalizing the rankings of sites violating Google’s existing Quality Guidelines. Violations specifically mentioned include keyword stuffing, article spinning and dubious linking habits like participation in reciprocal or paid link networks. In Google’s official Penguin release, Google goes on to point out that, downstream, this change will only impact around 3% of total, US queries – they’re looking for the worst of the worst offenders. (Compare this to the 35% change that last year’s QDF update brought to bear.)

Because of the extreme nature of the violations Google is targeting, most businesses with a sound marketing team will not see any penalty from the Penguin update. It’s difficult to imagine this statement is true, given the buzz on the web about the negative impact of Penguin, but keep in mind most of the “sky is falling” comments online are coming from novice-oriented forums. Yes, there have been a few mistakes where Google “overreached” in their penalty but, largely, the forum world has formed an echo chamber, amplifying the struggles of those who’ve ignored SEO guidelines for many years. The short story: if your content reads well (to a human) and you haven’t been buying/selling or exchanging links with suspicious sites, Penguin will not impact your rankings.

How can you tell if Penguin has been assessed against your website? Well, Google has been sending out warning messages about bad link networks for some time but there is no Webmaster Tools message specifically designed to tell you if Penguin has been brought to bear. The best advice is to check your traffic on and after April 24th – if your traffic went up or was unchanged, you can be fairly certain you didn’t fall victim to this ranking shuffle.

If your site traffic went down appreciably after April 24th, Penguin may have hurt your rankings. To recover, clean up the offending content and stop participating in paid and other high-volume link schemes. Unfortunately, a reconsideration request through Google Webmaster Tools is not the next step – Google has already said “Because this is an algorithmic change, Google has no plans to make manual exceptions.” This leaves two options for recovery:

i)for those sites truly penalized by mistake, owners can submit feedback through Google’s Webmaster Forum or submit this form to report the change as an error

ii) for those sites that have been penalized and subsequently cleaned up their over-optimization efforts, the best advice is to wait 3-8 weeks. If Penguin will be run intermittently (like “Panda“), the next run should happen within that time period. If Penguin, instead, is an in-residence part of Googles’ algorithm (as I think it is) waiting the same period of time is worthwhile but chances of recovery are slim.

At the end of the day, the Penguin panic is extremely overblown. The over-optimization “rules” recently enforced have been around for 10+ years as well-known SEO best practices not to mention having been included in Google’s Quality Guidelines. If your site was penalized and that penalty was not an error, good advice would be to clean up your site’s footprint but, moreover, take a long look at who is executing your search marketing… they should have known better.

The news that the US government has hired an outside prosecutor to represent it in an antitrust inquiry against Google is ratcheting up the discussion of whether and how much Google abuses its power. Google has 66% market share for search and is the largest traffic-driving source for most online marketers. The heart of the issue is whether or not Google is giving preferential placement on the SERP to Google-owned properties, ranking them above their competitors. This action by the government indicates as serious a commitment to investigate as it did with Microsoft 14 years ago, and there has not been as high profile an antitrust case since.

What the government is focusing on is worthy of looking deeper at;  however, there are other issues with Google that we all know about that are questionable, too.  While we may have become accustomed to them, they remain egregious.  In our own dealings with the myriad of people we work with at Google, we regularly express our dissatisfaction in some or all of these areas.  Our agency’s feedback to Google has helped create some significant changes over the years.  As such, we continue to push where we can.

1.  Black box pricing in AdWords – Nobody knows how the starting price of the auction is determined. There is much cynicism in the industry about Google giving themselves a raise, when needed, by systematically increasing bids by a penny or so.

2.  Black box ad placement with AdWords - Search marketers understand the pillars of Quality Score (bid, relevancy and CTR) and how Quality Score affects placement of paid ads. However, there isn’t an empirical formula that maps it out, not to mention that determining “relevancy” is completely subjective.

3.  Continual changes in the natural search algorithm that often catch marketers off guard – Panda caused a significant drop in traffic to many websites. Lots of website owners were unprepared since the content guideline with Panda was different than we’d seen with other algorithm changes in the past (focusing more on quality of content vs. spam). The SEOs benefited the most from this since many had to seek professional help to undo the damage.

4.  Random determination of what is considered spam - Overstock lost visibility on the SERP last year for soliciting students to link to Overstock’s site on .edu sites. It is arguably a gray area, but Google considered it spam, anyway, and manually delisted them. Overstock saw its natural search traffic disappear for a period of time. Now we are hearing that “over optimized” sites, another subjective determination, may soon be considered spam, too. There have been others in the past and will be more in the future. Brands need to remain particularly vigilant over who and how their SEO is being managed. On Google’s end, they should not just take down a site that they feel is spamming. Rather, they should give the webmaster a warning, and the time needed to fix the issue. E-commerce is very big business now. A month, a week or even one day down for a retailer can create very big losses.

5.  Enabling anyone to bid on any brand’s trademark - Advertisers must bid on their own trademark or risk losing traffic to a competitor. Even if the brand already has top placement in natural search, with the paid ad being higher on the page and shaded, competitors can siphon off a good chunk of traffic if they run on someone else’s brand term if the true brand owner itself is not in Position 1 in paid search. Interestingly, the only trademark term Google does not let anyone bid on is “Google” itself.

Since I live and breathe attribution and optimization, I thought I might try clarify what they are and why they are important.

In regard to attribution, I am continually surprised at how many direct marketing professionals don’t understand what it means. I recently saw the word “attribution” on a buzzword bingo card at a trade show. It is not a buzzword! Or a fad. It is critically important to the health of a business.

Attribution has to do with how revenue is allocated across sales channels (either in the case of just online or offline to online) or keywords (in the case of paid search) or devices. Once the revenue is allocated appropriately, the marketer can calculate an ROI to understand the profitability and true contribution of each sales channel. This information informs investment decisions.

Attribution modeling is necessary in the online channel because cookie lengths and multiple tracking pixels cause revenue/orders/sales to overlap across marketing programs. When a marketer receives their performance reports from their paid search, email, affiliate, display, CSE, and remarketing vendors, the duplication is evident insofar as the total of all these programs doesn’t match what is reported in web analytics. The overlap can be substantial (20% incremental would be considered on the low side, for example). Ultimately, the number needs to be 100% — not 120% — so decisions need to be made about which marketing programs will win credit for the overlapped sales. Once the orders are de-duplicated, ROI can be calculated at the marketing program level. This is typically the circumstance that drives the attribution discussions.

There is also keyword attribution that occurs specifically in the paid search source. There are consumers who conduct a single search on a search engine, click on a link and then either convert or not convert. In this case, the ROI is a simple calculation of revenue into cost per click. There is another set of consumers (at least 30%) who conduct multiple searches before converting. Depending on the category of merchandise or type of action, it could be much more (for example a high ticket item will result in more multi clicks). In the multi click scenario, someone might search for a generic item such as “laptop’” and then search on — and subsequently convert on — the brand term “Dell.” The keyword “laptop” assisted in the conversion on the brand term and it could be argued that it should get some of the revenue credited it to it. If not, that keyword on its own will be unprofitable but the sale on the brand term never would have happened without it. Advertisers could choose to either split the revenue across all the keywords in the clickstream, give it all to the assist term (laptop) or give it all to the trademark term (Dell). Whatever model the advertiser chooses is at their discretion. However, many search marketers are still constrained by the technology they are using to manage their paid search with the majority in the industry having no choice but to go with either last or first click, depending on which platform they are on. While PM Digital can offer a choice of models to its clients, for many other advertisers, flexibility here is still a ways off, but it’s coming.

There are other forms of attribution that marketers are grappling with, too. An emerging discussion is device attribution in which a user initiates a search on a website from multiple devices. At present, cookie synchronization does not exist, so there is no way to track the same user from device to device. Yahoo is working on a tracking mechanism for this if the user is signed into their Yahoo account on all devices. We suspect Google is doing the same. We will cover this in more depth in an upcoming post.

I will address optimization in an upcoming post, as well, but just to clarify how it differs from attribution, they are really not related in any way. Optimization as it relates to paid search entails making changes in the search engine interface (or through an editor) to improve a campaign’s performance. These changes could include keywords, landing pages, match types and negatives, campaign structure, copy, geo targeting, day parting, bidding and more.

Suzy Sandberg is President of PM Digital.

• Misspellings (“restaurant” & “restaraunt”)
• Singular/Plural Forms (“restaurant” & “restaurants”)
• Stemming (“dine” & “dining”)
• Accents (“café” & “cafe)
• Abbreviations (“mr. chow restaurant” & mister chow restaurant”)
• Acronyms (“la restaurant” & “los angeles restaurant”)

At least 7% of search queries on Google contain a misspelling, and the longer the query the more likely that is to happen. Google’s organic search already detects and compensates for misspellings and close variants by showing results for the user’s intent (see below).

Google believes this new feature will be beneficial for advertisers and will help result in more impressions and clicks for phrase and exact match. Early experiments show that advertisers experienced positive results with an increase in clicks by 3% with comparable CPCs.

As we’re relying more heavily on exact and phrase match to improve ROI, the misspellings and other variants have been a good opportunity for keyword expansion for those advertisers who diligently work through search query reports.

While this change in match types can help save time in keyword development overall, it may result in a more crowded landscape on these variants. Quality Score and first page bid estimate will be determined on the exact keyword according to Google, but CPCs could also rise if advertisers bid up the main phrase and exact match keywords.

Advertisers can opt out of this feature only at the campaign level by selecting keyword matching options under the advanced settings.

Multi-channel marketing is mostly still multiple channels operating simultaneously. While it may provide some brand consistency, it is not fully integrated within and between channels or fully embedded in the brand voice or offer copy. With all channels now having the technology to support an integrated approach, there is significant potential to leverage marketing dollars.

For instance, with all of the angst around measuring the impact of social media as a stand-alone channel, we can now incorporate social media messaging and offers into email, in real time. Nothing has quite the cultivation power as email, and its performance is easily measured. So, to extend this example, if, in an email, you pushed your social media as a destination for contests which provides all of the rich context of other participants, but the entrant had to “email to win”, you would get a very different measurement of social media’s contribution.

Not all of the measurement systems are fluid or tied together, but it’s coming, and how it impacts decisioning around marketing and how dollars get spent is exhilarating…and an outright gamechanger.  Check out the Winterberry Group whitepaper entitled Rethinking the “Four Ps”: Marketing Operations Management and the New Pathway to Productivity.

What is SSL search?

In October 2011, Google announced that it was automatically redirecting signed-in users to the https version of the Google search site. All search traffic sent through this URL will be encrypted via SSL as it leaves the searcher’s browser and goes through their router to their ISP, Google and then on to Google’s recommended websites. The primary reason is to protect the privacy of users’ searches if they are conducted via unsecure WiFi or public networks. I would offer two other possible reasons for this move: i) limiting liability; and ii) enhance the value proposition of PPC.

Limiting liability – organizations in both the US and EU have challenged Google’s privacy policies, particularly data sharing and user protection elements. While I’m not a lawyer, I can imagine Google’s liability would be significantly reduced if they encrypt search data for signed in users – users who have, further, agreed to terms of service that can dictate additional protections for Google. Is this speculation? Absolutely… but, for Google, winning at this level means focusing on lawsuits and fighting patent disputes, not brand development through feel-good open data policies.

Value proposition of PPC – if we view websites participating in Google’s AdWords & AdSense networks as legally-contracted extensions of Google, we could hypothesize that this could provide the legal protection needed to allow Google to continue to offer keyword level data to participants. Happily for Google, this also enhances the value proposition of participation in those PPC networks, as it will become the best way to get keyword level data. (Google Webmaster Central still shows the search terms people used to reach your site over the past 30 days but it is not as integrated as PPC data is or would be.)

What’s the impact of encrypted search on site data?

If you depend on Google organic results for any significant portion of your website traffic, you’ve seen the impact of encrypted search within your site analytics as keyword “not provided” within your referrer reports. In the first few months of SSL encryption, surveys showed only around 9% of traffic data being impacted. Since then, voices ranging from Baynote.com to Danny Sullivan have joined the conversation and noted up to 35-37% of organic search data being impacted. Based on my conversations with retail clients, 15-20% of organic search data is being returned as “not provided” for that vertical.

Cue the ominous music… days ago, it was revealed that the newest version of Firefox will also use Google’s SSL-encrypted search as its default. SearchEngineLand.com does a great job explaining the details but this certainly means another 5-20% of organic search data will be rendered not available, depending on the percent of Firefox users and early adopters within your site’s visitor profile.

How to manage?

So what’s a data-driven webmaster to do now that 10-35% of their traffic is being returned as “(not provided)”? The answer is, largely, to imply patterns for the (not provided) searches using keyword data you do have. There are a few interesting considerations brought up by Paul Burani at SearchEngineLand.com, though. In his article Google’s Encrypted Search Data: A Cure for Vision Loss?, Burani notes that searchers who are signed in to Google are more likely to be

• More familiar with the website (and possibly the brand itself).
• More likely to reside in North America.
• More likely to land on a deep URL after a search query (instead of the home page).

Use these considerations to tweak your assumptions, slightly, toward brand-related searches by US visitors. Further, filling the data gap can also be addressed using other methods, including:

1. connecting Google Analytics to your Google Webmaster Central account for better Top Query keyword reports;
2. looking at non-keyword-related visitor data (return vs first time visits, time on site, top landing pages, etc.);
3. data mining your own site’s search function;
4. re-considering the importance of Bing and Yahoo! search traffic data.

It’s an unavoidable fact that fully integrated, organic keyword referral data is going to become increasingly scarce. Regardless of underlying factors, that’s likely for the best for end users’ privacy. From a business perspective, for now, the problem is limited to a minority of searches. Assuming the trend continues and a service like Google+ doesn’t arise to provide a legal way around the issue, search marketing pros and third party data providers are likely to continue scrambling for a solution.

The core idea is that low quality content is being carefully inspected and, if you go over a certain threshold, your entire site can be penalized. There are ways out of the Panda hole but what do you think, generally, about this series of updates? Has Panda hit your site? Is it helpful to retailers or harmful?

Mobile searches, including tablets and smart phones, have grown 5 times in the past 2 years, and one in every three Americans will own a tablet by 2015. (According to Google data)

• 82% of users of mobile devices are researching, which includes looking for product information, etc. (According to Google’s Mobile Movement strategy)

• 68% of users visit a business after conducting a search on mobile while 53% of users purchase a product. (According to Google’s Mobile Movement strategy)

• 1 in 3 shoppers use Smartphone’s (most often a search engine) to find a store location and nearly 20% of users have looked for a coupon. (comScore)

• More than 20% of Smartphone users contacted friends or family while in a retail store to discuss a product, and 12% compared prices from other retailers. (comScore)

• Mobile spending peaked in the 3rd and 4th quarters of 2011 after the iPad 2 was released last Spring. 9% of online sales came from a tablet or mobile phone in Q4, up from just 3% from 2010. (comScore)

Best Practices – Mobile Search

The question on every advertiser’s mind is how to make the most of their Mobile opportunity. We’ve devised a list of best practices to help you sort through the wealth of information available.

1. Mobile and Tablet campaigns should be broken out from traditional search campaigns as opposed to targeting all devices. This allows the advertiser to:

• Customize Landing Page by Device
• Use Mobile or Tablet Specific Copy
• Review KPIs for Each Device Separately
• Set Bids by Device as the Search Landscape Varies
• Budget Separately

2. Mobile Sitelinks can be used to help improve CTR within search ads and can also help drive visitors further into a section of the site that they’re looking for, which may be difficult to navigate to on the device.

3. For some users who are searching on a smartphone it might be best to offer them a phone # where they can call for more info, customer service or to purchase. Google’s Click to Call Extensions make this easy.

4. In addition to Sitelinks, Google also offers Mobile App Extensions where advertisers can increase the distribution and download of any apps relevant to the search. The ad includes the app rating, price (if applicable) and links users directly to Android Marketplace or the iTunes App Store.

Best Practices – Tablet

While some of the same advantages highlighted on the mobile slide apply to Tablets, these device campaigns tend to perform more like computers. For eCommerce, tablets regularly convert at the same rate and often produce higher average order values.

The most important consideration is the ad placement and bidding to ensure adequate coverage. Up to three ads can show on the top while three ads can also show on the bottom. There are no ads down the side.

• Sitelink Extensions are Available
• Target Top 3 Positions for Ads
• Consider Bid Boosting During the Evening for Peak Tablet Usage
• Test Tablet-Specific Ad Copy or Promotions
• Create New Landing Pages if Those Used for Desktop & Laptops are Not Fully Compatible with iPad or Contain Extensive use of Flash
• Track Separately for Performance and Bid Management

Mobile Advertising

While mobile search ads are the best options for driving leads and conversions, AdMob and other networks can be used for branding, building awareness and promoting apps.

Demographic and specific mobile provider-targeting is often available, and ad formats range from simple banners to rich media and video. If app distribution is the name of the game, you can use text or banner ads to link directly to an app store for easy download.

If you have questions, or would like to get started right away, contact us info@pmdigital.com.