Brand Equity & Forcing the Brand Buy

Yahoo! put out a research paper highlighting activity bias, stating that the efficacy of online advertising is often over-stated because people who see ads about a topic were already more closely tied in with that particular network & that particular topic before they even saw the ad. As an example, any person who sees an AdWords ad for hemorrhoid treatment was already searching for hemorrhoid-related topics before they saw your ad (thus they were in the subset of individuals that might have came across your site in some way if you were in the search ad ecosystem or not). This sort of activity bias-driven selection bias (homophily) exists on social networks online & offline. Google did research on incrementality of ads & they came to the opposite conclusion as Yahoo! did. Google suggested you should buy, buy, buy, even on your own branded keywords. They suggested that testing was expensive (no mention that the only reason it is expensive is because Google chooses not to make such tools easily accessible to advertisers) & that the clicks were so cheap on branded keywords that you should buy, buy, buy. Many advertisers who mix brand & non-brand keywords together don't realize that they are using the "returns" from bidding on their own brand to subsidize over-paying for other keywords. Google Analytics is the leading & most widely used web analytics program. They can share whatever metrics help them sell more ads (defaulting to crediting the last click for conversions, even if it was on a navigational search to your site) & pull back on features that are not aligned with their business interests (SEO referral data anyone?) This goes back to Scott McNealy's quote: "The only technology I’d rather own than Windows would be English. All of those who use English would have to pay me a couple hundred dollars a year just for the right to speak English. And then I can charge you upgrades when I add new alphabet characters like ‘n’ and ‘t.’ It would be a wonderful business."

Thin Content & Scraper Sites

Remember the whole justification for Panda was that thin content was a poor user experience? In spite of sites like eHow getting hit, Google is still pre-paying them to upload content to Youtube. Now that the (non-Google hosted) thin content has been disappeared (and the % of downstream traffic from Google to Youtube has more then doubled in the past year) it is time for Google to take another slice of the search traffic stream with Search Plus Your World: The Google vs Facebook locked down walled garden contest will retard innovation. As the corporate internet silos grow larger the independent web withers. Them going after each other may leave room for Twitter, but it doesn't leave lots of room is left for others, as the economics of publishing have to work or the publishers die. Start ups that were on a successful trajectory were killed by Panda: The startup had been on a roll up until last February when Google altered its ranking algorithm with the release of “Panda.” The changes decimated TeachStreet’s traffic, and the company never quite recovered. “We lost a lot of our traffic, and overnight we started talking to partners for biz dev, not for acquisition,” he said. However, many of the potential partners wanted to know about an outright acquisition. About.com was also smoked by Google: The biggest worry, though, is that the decline of About.com itself may be irreversible. Fewer people are clicking on About ads placed by Google and the site’s own display ads have dropped in value. The company has attributed this decline in value to Google’s decision last year to downgrade About pages in its search results. With more than 80% of traffic coming from search, the Google denigration was indeed a blow but About’s problems may be rooted in something deeper.

Web Scrape Plus+ (Now With More Scraping)

When the +1 button & Google+ launched, Google highlighted how they would use the + button usage as a "relevancy" signal. Google recently started inserting + pages directly into the search results for brands & right from the very start they were using it as a scraper website that would outrank the original content source. Google used the buy in from their promised relevancy signal to create a badge-based incentivized system which acts as a glorified PageRank funnel to further juice the rankings of these new pages on a domain name that already had a PageRank 10. I recently read a blog post about how anyone could do the above & the opportunity is open to everyone. But the truth is, I can't state that something will become a relevancy signal that manipulates the search results in order to get buy in. Or, if I did something which actually had the same net effect, Google would likely chop my legs off for promoting a link scheme. Recently the topic of Google+ as a scraper site came up yet again via Read Write Web & on Hacker News a Googler stated that it was "childish" to place any of the blame on Google!!!!!! Google determines how much information is shown near each listing & can create "relevancy" signals in ways that things tied to Google get over-represented (look at the +1 count here). When they do that & it destroys other business models *of course* Google deserves 100% of the blame.

Bogus Testimonials & Social Payola

Is social media a cleaner signal than links? If search engines put the same weight on social media that they put on links it would get spammed to bits. It won't be long until a firm like Ad.ly offers sponsored Google+ posts. Some have suggested that you won't be able to buy Google+ followers however Google already includes user pictures on AdWords ads (even when they desire not to be & even when they didn't endorse the product that Google suggests they endorsed). In due time I expect Google will indeed sell followers & other user interactions as ad units (just like Twitter & Facebook do). Further, celebrities sell Tweets to advertisers. When they are hot their rates go up: When Ad.ly introduced self-destructing Charlie Sheen to Twitter, he was paid about $50,000 per tweet. It was worth it. Sheen’s tweet for Internships.com generated 95,333 clicks in the first hour and 450,000 clicks in 48 hours, created a worldwide trending topic out of #tigerbloodintern, attracted 82,148 internship applications from 181 countries, and added 1 million additional visits to Internships.com. Search engines might consider these to be clean signals if those same search engines were not busy buying the manipulation of said "relevancy" signals. Attention is purchased to create demand. It isn't comfortable to put it this way, but we are trained to obey authority & to like what others like: The average Facebook user has 130 friends, which equates with four degrees of separation to thousands of people, Mr. Fischer said. Metrics like that led him to believe that if Facebook could figure out a way to capitalize on "social endorsements," it would be like creating a word-of-mouth campaign that could reach millions of people simultaneously. Since the campaigns would come from a friend, they would theoretically be taken more seriously than, say, a TV commercial, he said. On an individual basis reviews and ratings get faked everywhere. Even stodgy old slow-moving institutions like colleges game their ranking systems. There recently was a question raised about how Google's rating systems skewed high on the underlying data. Surely Overstock (the same Overstock Google penalized earlier this year) wouldn't promote Google's trusted stores aggressively on their own site if it made their business appear worse than it actually is, thus a positive bias must be baked in to the system.

User Privacy

Ahead of Google updating their privacy policy Google has directed a large portion of their ad budget toward ads about how they protect users online. What better way to ensure user privacy than to allow them to register their accounts under psydonyms? The real name policy on Google+ was part of what made Google want to stop providing referrer data for logged in users who search on Google. This has had a knock on effect where other social sites are framing everything, requiring registration to read more of public user generated content & sending outbound traffic through redirects. Google's new privacy policy allows them to blend your user data from one service into refining the experience (and ads) on another: If you’re signed into Google, we can do things like suggest search queries – or tailor your search results – based on the interests you’ve expressed in Google+, Gmail, and YouTube. We’ll better understand which version of Pink or Jaguar you’re searching for and get you those results faster. Google & Facebook's war (against) user privacy is catching media and governmental attention. Microsoft highlighted some of Google's issues in their "putting people first" ad campaign & the blowback has caused Google not only to publish PR-spin "get the facts" styled blog posts, but to launch yet another ad campaign.