By Jeff Watts, SyndicateVoice Product Manager

Question: What do the following silly things have in common?

1.    what is the approximate size of a banjo
2.    redheads and the men that love them
3.    how to hit the sasquatch
4.    used men’s hunting socks

Answer: they are all actual phrases that searchers have typed in to discover various Bazaarvoice clients’ review content.  (For the record, I cannot fathom why someone was searching for “used men’s hunting socks”.)

The long tail of search engine data tells a fascinating story that smart marketers are listening to.  It is not just silly phrases like those above, but obscure and unique terms – perhaps used only once or twice – that give insight into what your visitors are really looking for.  Long tail search terms might not match your page titles or your carefully crafted page descriptions, but they do match some other combination of words on your page – even when you are not anticipating them.   For many sites, the unanticipated terms account for over 80% of the search terms and over 50% of the referrals, and this makes it imperative to understand how best to target those terms.

With consumer-generated content, targeting unanticipated search terms is easy.  Consumers writing reviews, for example, are using their own words to describe their experiences with a product or service.  Their reviews do not go through a formal editorial process to correct misspellings or to remove personal words like “I”, “me”, and “my”.  The reviewers themselves are not employees who must adhere to the official company marketing message in what they write.  Further, a typical page of reviews is written by multiple, unrelated authors who have different perspectives on the product being reviewed, and these different perspectives result in a greater variance in the words used in the reviews.  This variance means that pages of reviews are more likely than homogenous product pages written by one or two authors to drive search terms from unanticipated combinations of words, that is, long tail search terms.

For example consider the following excerpt of search terms from one of our clients’ long tail data.  Each of the following words was searched on just a few times during the reporting period.  Notice the variance in the words and phrases used:

•    nike free trainer 7.0 review
•    nautica jeans colorblocked polar fleece pullover
•    love tease pleated rhinestone-trim dress
•    desperate housewives perfume
•    simmons beautyrest adrian
•    dz7064
•    love tease beaded organza halter dress
•    alfani safari field jacket
•    belique cookware
•    arturo chiang umiko
•    nike free 7.0 review
•    tools of the trade cookware reviews
•    dkny draped swimdress
•    reviews on benefit firmology
•    towelbuyer
•    love tease beaded babydoll dress
•    nike free trainer 7.0 reviews
•    keds mosaic slipon

Lists of long tail terms like this are typical for publishers of reviews:  some are product names, some include keywords like ‘review’ or ‘reviews’, and still others are brand names or generic descriptions of products.  Even though each term is searched on once or twice per reporting period, the sum of all of them together frequently accounts for tens of thousands or hundreds of thousands of search referrals.

No marketer can predict all of the terms that their searchers will use during a reporting period, but it is possible to identify the types of words that are most frequently associated with success – the words that are used most frequently in search referrals that lead to conversions.  For example, consider the following illustration.  On the left side is the excerpt of long tail data mentioned above with each keyword phrase now associated with the number of “Referrals” generated and the number of “Conversions” accomplished during the reporting period.  (A referral is counted when a searcher clicks through to your page from the search engine.  A conversion is counted when the visitor subsequently takes the desired action on your site, whether that is a purchase or some other metric.)
The list on the left is aggregated into a similar list on the right by tallying the number of conversions that were associated with each individual word in the search phrases on the left.  For example, the words “nike”, “free”, “trainer”, “7.0”, and “review” each count for one conversion, based on the first row of data in the chart on the left.  Tallying these numbers up for the entire list (including the rest of the search terms that were not included in the list above), we see some interesting trends emerge.  For one thing, the most important words for generating conversions were “review” and “reviews”.  This is not unexpected, since we are dealing with pages that contain review data.  Other words that led to conversions were “top”, “best”, “ratings”, and so on.

Wise marketers use lists like this to help determine which keywords to put into the page templates of their consumer-generated content pages.  (Incidentally, this is one of the basic principles Bazaarvoice uses in building the SearchVoice Pages that come with our Ratings and Reviews product and in building the SyndicateVoice microsite.)

The long tail is a gold mine for search marketers and the variance of the content contributed by consumers helps you tap it.  Read and analyze terms in the long tail of your data to find out what your site visitors are really looking for.  Oh, and if you are the one that was looking for used men’s hunting socks, please drop me a line and let me know what you were thinking.

6 Responses to “The Long Tail Opportunity of Consumer-Generated Content”

  1. Julian Chu

    Great article, Jeff! How can one easily compile the aggregated per-word conversion count that you use as an example in this article? Is there an easy Excel trick that will do it? Appreciate any help you can provide on this. Feel free to contact me directly.

    Thanks,

    Julian Chu
    Director, Client Success
    Demandware, Inc.

  2. Julian Chu

    Great article, Jeff! How can one easily compile the aggregated per-word conversion count that you use as an example in this article? Is there an easy Excel trick that will do it? Appreciate any help you can provide on this. Feel free to contact me directly.

    Thanks,

    Julian Chu
    Director, Client Success
    Demandware, Inc.

Leave a Reply

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>