Analytics & Data

SEO: Measuring Performance of the Long Tail

The long tail should drive most of the natural search performance for ecommerce companies. But whether it actually does is another matter. Don’t assume that the long tail is working for you — make sure that it is, or fix it.

This is the third installment in my three-part series on addressing the long tail for search engine optimization. In the first installment, “SEO: How to Maximize the Long Tail,” I explained the basics: the propensity for the individually high-value trophy keyword phrases to drive less natural search performance in aggregate than the phrases that drive only a few searches each.

The second installment, “SEO: Why Your Long Tail Isn’t Long,” covered strategic, architectural, and technology challenges that can cut the long tail short of its full potential.

The long tail of marketing also applies to SEO. Image:

The long tail of marketing also applies to search engine optimization. Image:

Long tail theory assigns the labels “head” and “tail” to groups of keywords, as shown in the graphic above. The “head” of the curve represents the keywords that drive the most traffic individually. If you sell toys, those head keywords would be words like “toys” and “games” and your company’s or site’s brand name. The “tail” of the curve is everything that doesn’t drive large amounts of traffic for each individual keyword — keywords like “wooden puzzles for adults” and “pink Monster High vampire doll.”

Analyzing Keyword Performance

But here’s the reality. The only way to determine whether the long tail is working for your ecommerce site is to compare your keyword research with your analytics: matching the keyword report from your analytics with the keyword research you’ve already done.

In this way you could see how many times a month on average a word or phrase has been searched for and also the number of visits your site has earned for that same keyword. Applying a VLOOKUP formula in Excel can manage that task handily. See Microsoft’s support site for more on the VLOOKUP formula. My article on using Excel for SEO is dated, but still applicable, at “SEO: Using Excel for Keyword Analysis.”

Unfortunately, since most keyword data is not available, we can’t trust that data in our own analytics tools. We need to use a proxy. Google Search Console (formerly, Webmaster Tools) shows, in its Search Analytics report, the top 999 keyword queries that drove traffic via Google in any time period in the last three months. The data is also available in Google Analytics if you have linked the two tools together.

The Search Analytics method has a couple of limitations. First, the data is for Google only. You won’t see how the tail is performing in Bing, Yahoo, or any other search engine.

Second, the time frame is only the last three months, which means you can’t analyze your performance against past performance unless you’ve been regularly saving the data to your hard drive (something I recommend).

And third, unless your site has a very limited set of products for sale, your long tail will almost certainly be longer than 999 keywords long. If not, that’s the first indicator of a problem.

Analyzing Entry Page Performance

The long tail of search also corresponds roughly to different page types, mapping to the typical hierarchy of an ecommerce site. The head phrases at the front of the long tail curve — like “toys” and “dolls,” which drive a lot of keyword demand individually — would typically be targeted by the home page and major category pages of a site selling toys. And the long tail phrases like “wooden puzzles for adults” and “pink Monster High vampire doll” would be targeted by some of the faceted subcategory pages and product pages, which are typically deeper in the site.

The long tail of search also corresponds roughly to different page types, mapping to the typical hierarchy of an ecommerce site.

A few pages at the top of the site target the few head phrases at the front of the curve, and the many pages at the bottom of the site target the many keyword phrases found in the long tail of the curve.

SEO professionals also add in the concept of the chunky middle. The keywords that each drive decent traffic individually — like “wooden puzzles” without the “for adults” modifier — but aren’t large enough individually to be considered head phrases fit into this category. Chunky middle keywords correspond to lesser category pages, subcategory pages, and some facet pages.

With that mapping in mind, we can use as a proxy for long tail performance a natural search entry page report in our analytics. This works best for the more targeted pages deeper in the site — the pages representing the chunky middle and the long tail.

For example, if your natural search entry page report shows that the home page and a handful of main category pages are driving all of your natural search traffic, you likely have a problem with your chunky middle and long tail, as the pages that would target chunky middle and long tail phrases are not performing.

Similarly, if only your home page, a handful of main category pages, and your product pages are performing, you’re likely missing your chunky middle. A missing chunky middle is more common than you’d think.

The other common issue is a chunky middle hidden within the long tail. In this case, the pages that should be driving moderate amounts of traffic are instead driving very small amounts. They will be hard to identify within the sea of long tail-targeting pages like product pages.

What should you do when you’ve discovered your poor long tail performance? Start with the second installment in this series, “SEO: Why Your Long Tail Isn’t Long.”

Jill Kocher Brown
Jill Kocher Brown
Bio   •   RSS Feed