• Home
• About
• Writers
• Landscape
• Wizards Index
• Fraud
• Wanna Be Like Larry?

Exogenous Complexity 4: SEO and Big Data

Introduction

In the interview with Dr. Linda McIsaac, founder of Xyte, Inc., I learned that new analytic methods reveal high-value insights about human behavior. You can read the full interview in my Search Wizards Speak series at this link. The method involves an approach called Xyting and sophisticated analytic methods.

One example of the type of data which emerge from the Xyte method are these insights about Facebook users:

• Consumers who are most in tune with the written word are more likely to use Facebook. These consumers are the most frequent Internet users and use Facebook primarily to communicate with friends and connect with family.
• They like to keep their information up-to-date, meet new people, share photos, follow celebrities, share concerns, and solve people problems.
• They like to learn about and share experiences about new products. Advertisers should key in on this important segment because they are early adopters. They lead trends and influence others.
• The population segment that most frequents Facebook has a number of characteristics; for example, showing great compassion for others, wanting to be emotionally connected with others, having a natural intuition about people and how to relate to them, adapting well to change, embracing technology such as the Internet, and enjoying gossip and messages delivered in story form and liking to read and write.
• Facebook constituents are emotional, idealistic and romantic, yet can rationalize through situations. Many do not need concrete examples in order to comprehend new ideas.

I am not into social networks. Sure, some of our for-free content is available via social media channels, but where I live in rural Kentucky yelling down the hollow works quite well.

I read “How The Era Of ‘Big-Data’ Is Changing The Practice Of Online Marketing” and came away confused. You should work through the text, graphs, charts, and lingo yourself. I got a headache because most of the data struck me as slightly off center from what an outfit like Xyte has developed. More about this difference in a moment.

The thrust of the argument is that “big data” is now available to those who would generate traffic to client Web sites. Big data is described as “a torrent of digital data.” The author continues:

large sets of data that, when mined, could reveal insight about online marketing efforts. This includes data such as search rankings, site visits, SERPs and click-data.  In the SEO realm alone at Conductor, for example, we collect tens of terabytes of search data for enterprise search marketers every month.

Like most SEO baloney, there are touchstones and jargon aplenty. For example, SERP, click data, enterprise search, and others. The intent is to suggest that one can pay a company to analyze big data and generate insights. The insights can be used to produce traffic to a Web page, make sales, or produce leads which can become sales. In a lousy business environment, such promises appeal to some people. Like most search engine optimization pitches, the desperate marketer may embrace the latest and greatest pitch. Little wonder there are growing numbers of unemployed professionals who failed to deliver the sales their employer wanted. The notion of desperation marketing fosters a services business who can assert to deliver sales and presumably job security for those who hire the SEO “experts.” I am okay with this type of business, and I am indifferent to the hollowness of the claims.

What interests me is this statement:

From our vantage point at Conductor, the move to the era of big data has been catalyzed by several distinct occurrences:

• Move to Thousands of Keywords: The old days of SEO involved tracking your top fifty keywords. Today, enterprise marketers are tracking up to thousands of keywords as the online landscape becomes increasingly competitive, marketers advance down the maturity spectrum and they work to continuously expand their zone of coverage in search.

• Growing Digital Assets: A recent Conductor study showed universal search results are now present in 8 out of 10 high-volume searches. The prevalence of digital media assets (e.g. images, video, maps, shopping, PPC) in the SERPs require marketers to get innovative about their search strategy.

• Multiple Search Engines: Early days of SEO involved periodically tracking your rank on Google.  Today, marketers want to expand not just to Yahoo and Bing, but also to the dozens of search engines around the world as enterprise marketers expand their view to a global search presence.

All the above factors combined mean there are significant opportunities for an  increase in both the breadth and volume of data available to search professionals.

Effective communication, in my experience, is not measured in “thousands of key words”. The notion of expanding the “zone of coverage” means that meaning is diffused. Of course, the intent of the key words is not getting a point across. The goal is to get traffic, make sales. This is the 2112 equivalent of the old America Online carpet bombing of CD ROMs decades ago. Good business for CD ROM manufacturers, I might add. Erosion of meaning opens the door to some exogenous complexity excitement I assert.

The notion of a “universal search” is a weird one. Google whipped up the phrase to mask the fact that its indexing method worked on silos. Even today, merged and reduplicated results are not the company’s core competency. I find the admonition to “get innovative” interesting.

Finally, I puzzled over the reference to “dozens of search engines.” The data I have seen is that Google has the lion’s share of search traffic. It follows that if one is indexed in Google, game over. Furthermore, some of the “other search engines” make an effort to match Google results. Test this yourself. Run queries across Jike.com, Google.com, and Yandex.com. Look a the first page or two of results. How different are these services? In short, what turns up in Google often turns up in search engines elsewhere.

What does the article’s inclusion of these three points have to do with “big data”? I think, not much.

The Xyte Method vs. Random SEO Big Data

I want to point out that the Xyte method is more sharply focused. The methods described in the Search Engine Land write up are not. Xyte is just one of a number of companies with methods that uncover nuances in human behavior. The method is just that, methodical. The Search Engine Land article is more like a finger-painting done in haste.

The exogenous complexity comes into play when one uses either the Xyte-type approach or the SEO-type approach. But I think there is a difference: Xyte rests upon a specific analytic process. The SEO-process uses what’s available to look as if the approach is analytic.

In the real world, the SEO approach has devalued general Web index relevance. Google has failed in its efforts to maintain objective relevance because whatever actions it takes, the spoofers foul the engine. Google now “sells” traffic in the form of AdWords and, in my opinion, has continued to make research difficult for those who are looking for objective information. Now the SEO crowd is moving beyond Web indexes, seeking to twiddle the results in app store rankings, pizza joints displayed on maps, and Twitter follower counts.

But “big data” poses a problem. First, the cost of processing large flows of data can be high. In order to manage the costs, “big data”, like other types of data, get filtered or sampled. The problem is that the reliability of “big data” samples can be uncertain. The key to many “big data” analyses is identification of a meaningful signal or anomaly. In my experience, few marketing and advertising outfits possess the expertise to do this work. As a result, quantitative shops like Attensity and others provide a service. The outsourcing method means that the marketer accepts what the outsourced firm provides and may lack the expertise to ask the “right” question so that the limitations of a finding can be discerned. The approach guarantees that the client of the outsourcing firm passes the information to another person or company. With each step, the limitations get diluted. Some pretty crazy decisions result when “big data” turns out to be “big baloney.” How is that Web traffic doing? How are those Twitter followers turning into sales? Good questions.

The second problem is that search engine optimization has created a situation which produces predictable search results. Enter the word “gaga” into Google. You know the result list before you look at the results. Enter the query “definition gaga” and you know that the first hit will be to either an online dictionary or Wikipedia. The way Google works is that traffic determines results. Wonderful for popular queries but maddening for this query “shrink resistant steel.” The most relevant result is to a glove. The preferred result for me would be a link to Inconel 740. Search engine optimization has made it difficult to navigate a concept without detailed technical knowledge. I would argue that the more the SEO crowd fiddles, the more difficult finding valid information becomes. Look at the information Xyte surfaced about Facebook. The information is quite precise, has marketing relevance, and did not require gibberish about “big data.”

Finally, when findability becomes increasingly difficult, companies have a greater need to erode precision. That’s why the idea of thousands of key words is a problem for me. The steps taken to ensure findability erode relevancy the way concentrated hydrochloric acid nukes the finger of a careless first year chemistry student.

And Complexity?

I am reluctant to assert that SEO has dumbed down search and retrieval. I admit that the idea has a certain charm. I think that what started as tricks to make a Web site turn up at the top of a Google results list has made life more complicated for Google, researchers, marketers, and the hapless manufacturer of Inconel pipes for nuclear power generating plants.

Those first cheerleaders of SEO have complexified communication. Now an industry has arisen which is accelerating the erosion of meaning. We have zipped way beyond Orwell when a company has to think in terms of thousands of terms to describe Inconel. I thought “shrink resistant steel” was an okay phrase. I was expecting Inconel, but thanks to the incantations and own feathers of SEO experts with crazy data and buzzwords, Google fails me.

Who’s to blame? I think liberal graduates, former journalists, failed Webmasters, and confused political science majors lead the parade. Now that relevance is gone, marketers may have no prior warning that desperate measures are now common currency.

And objective search results? Forget that.

Stephen E Arnold, February 29, 2012

Sponsored by Pandia.com

Comments

Copyright © 2008 Beyond Search · ArnoldIT.com · Disclaimer · Revolution News theme by Brian Gardner · Log in