Dancing with Web stats

'Click' author shares search-behavior secrets

EMBRACE. FOCUS. EXECUTE. Build your 2019 roadmap to success with 4,000+ real estate leaders.
Inman Connect New York | January 29 - February 1, 2019

While there can be wisdom in crowds, the masses sometimes behave more like lemmings.

Take the buildup to the housing mess, for example: a lot of people making a lot of bad decisions.

As general manager of global research for online metrics company Hitwise, it is Bill Tancer’s business to glean useful information from online behavior — and to understand which information isn’t so useful and why.

Author of "Click: What Millions of People Are Doing Online and Why it Matters," Tancer is an expert in marketing, market research and corporate strategy who has previously worked for LookSmart, Zaplet, NBC Internet and Pacific Bell Internet Services.

During a keynote address at the Inman News Real Estate Connect conference in San Francisco, which runs from Aug. 5-7, Tancer will discuss what we can learn from online behavior.

Tancer has used online search statistics to help predict the outcome of reality television competitions such as "Dancing with the Stars" and "American Idol."

In a 2006 blog post, Tancer noted how many people were surprised at how "Dancing with the Stars" darling Stacy Keibler didn’t win the competition — especially because her online search popularity topped the charts compared to other competitors on the show.

Tancer later explained this paradox with the "Stacy Keibler Correction Coefficient": it wasn’t the volume of search traffic that was telling in this instance — it was the type of searches related to Keibler.

He noted in the blog that some of the top-searched sites related to Stacy Keibler were "visited primarily by 18- to 34-year-old males, not the most likely candidates to pick-up the phone and vote for their favorite ballroom dancer" … and the searches by this online audience didn’t always relate specifically to her dancing skills.

Tancer has also looked to Internet statistics as a possible predictor for trends in resale home sales.

"In today’s market, when you’re thinking about buying a house you’re going to turn to the Internet to look at listings and compare prices. If that holds, then the changes, flux in visits to the real estate category overall should be predictive of interest."

So as a commentator on CNBC, Tancer noted that he stood apart from other analysts during a news segment in predicting a monthly rise in resale sales stats based on a rise in online real estate search activity.

Again, the Stacy Keibler Correction Coefficient reared its head. "I was wrong," he said.

"We misinterpreted that interest in (the real estate) category as buyers’ interest." But some of the increased traffic was a result of consumers checking home valuations — not searching for for-sale homes. …CONTINUED

So there is still a ways to go to use real estate search statistics as a predictive tool for the housing market, he said. "There are a lot of pitfalls. We haven’t completely figured it out."

Hitwise has been studying search-term data to identify consumer confidence trends for several years, Tancer said, and search data did exhibit an erosion in consumer confidence from August 2007 to August 2008.

In a study of Internet users with household income above $250,000 per year, Hitwise found that this group tended to favor several finance- and investment-related sites in August 2007, such as Google Finance, Yahoo Finance and E-Trade.

But in August 2008, as the economic news ran dire, the search activity of this group changed dramatically, with entertainment- and game-related sites popping up among the most popular.

"You literally see (members of) this affluent segment put their heads in the sand, like they don’t want to know," Tancer said. In the following months, though, there appeared to be a renewed interest by this audience in financial matters.

Tancer notes that consumers’ real estate search behavior is very fluid, and it’s important for real estate professionals to understand this as they work to attract consumers to their Web sites.

The housing and economic downturn has rearranged consumers’ priorities and impacted search behavior. "You are constantly having to discover how events in the outside world (have impacted search)."

For search-engine-optimization purposes, the first few phrases in a blog post are vital in the post’s relevancy to searches, Tancer said. "If you want to talk about real estate in San Mateo Park, you want to put ‘San Mateo Park’ within those first 14 characters," he said.

Also, there is an increasing complexity in online searches. "People are getting very, very specific in their search queries. We have seen the five-plus-words-per-query that we track increase to between 5 and 8 percent of all searches."

And these longtail search phrases, because they are far more detailed, tend to be more valuable as leads than the less complex search terms, Tancer said. "You don’t want to miss out on this very specific high-converting traffic."

While studying your own site for clues as to how they got there can be valuable, Tancer said that macro-level data about the Internet community as a whole can reflect trends.

For example, he said that searches for the term "layaway" began to increase during this economic downturn, and retailers caught wind of this consumer interest and began to respond by launching layaway programs.

In addition to spotting retail trends, search data can also eliminate survey bias, as online actions speak louder than words. For example, Tancer said that consumer surveys about health-related fears can betray respondents’ true fears, while search behavior can be a better guide. …CONTINUED

When surveyed, consumers might express fears about serious medical conditions, for example, while there is a "completely different list" that is evidenced by search activity, such as "fear of other people" and "fear of being alone," he said.

Similarly, during speaking engagements Tancer has asked for a show of hands of those who view porn sites online — not surprisingly there are few if any hands raised. Search statistics can negate such selection bias.

When it comes to innovation, there are some clues to what makes a technology go viral.

Tancer said that search data has been used to examine visitor traffic demographics for the early days of online video site YouTube, for example.

YouTube visitor data for October 2005 showed some clear segmentation in the site’s audience, and Tancer said those same segments have appeared again and again at popular fledgling Internet sites.

He classified the visitor groups as "young digirati," "bohemian mix," and "money and brains."

"Those three groups were the common ones," and can serve as "a little crystal ball as to what may be hot tomorrow," Tancer said.

As consumers contribute more and more online content, that presents challenges in online searches.

"As a result of information getting very fragmented and spread out and thin, it’s going to cause search to be stretched," he said.

But therein lies an opportunity for companies "to leverage all of this information we provide about ourselves to provide much better results as we look for information."

There may also be a tug-of-war over how much privacy we’re willing to give up in exchange for searches that better fit our individual interests and profiles.

Bill Tancer, author of "Click: What Millions of People Are Doing Online and Why it Matters" and general manager of global research for online metrics company Hitwise, will deliver a keynote address during the Real Estate Connect conference in San Francisco, which runs from Aug. 5-7, 2009. Click here for details.

***

What’s your opinion? Leave your comments below or send a letter to the editor. To contact the writer, click the byline at the top of the story.