Amazon needs at least one visual representation of the long tail on its pages. It already uses the concept (in fact, the success of Amazon’s long-tail concept is almost the basis for the concept in the first place [see the third paragraph in Chris Anderson’s original article, “The Long Tail“) all throughout the site.

When you visit Amazon right now, variations of the long tail appear in Amazon’s linking of SIPs (statistically improbable phrases) and CAPs (capitalized phrases), its “Better Together” offer, its “Customers Who Bought This Book Also Bought,” its “Books on Related Topics,” its “Books That Cite This Book,” its “Customers Who Viewed This Book Also Viewed,” its “Explore Similar Items” (which expands the “Customers Who Viewed…” feature beyond the viewed item’s product category, e.g, for a book, what movies and music customers also viewed), its “Listmania,” its “So You’d Like To…,” its “Look For Similar Items By Category,” and, finally, its “Look For Similar Items By Subject.”

(of course, there is also the Amazon reader-reviews and forums, but to include those is to neglect what I think is the jazziest part of the long-tail concept: its technological dependence - the defense of which is way beyond the scope of this post)

To include a visualization of each of those variations of the long-tail would be gratuitous — I’m not even sure you could visualize it (as a long tail, anyway) for the “Books That Site This Book” variation. But it would be nice if Amazon did include a visualization of at least one variation.

Namely, the “Customers Who Bought This Book Also Bought…” variation. Right now, you get five text links on the product page, and three links (plus cover images) once you add the book to your shopping cart. These are great, but with the way Amazon has a tendency to include poorly related things on a page (for example, buying any TV series on DVD seems to mean that Amazon will constantly recommend — either passively or aggressively — any TV series on DVD), I would prefer to see the relationship between the original item and the item on the list.

If Amazon were to include a visualization of the long tail, I could see just how many customers who bought book X also bought book Y when compared to book Z.

How would this look? Well, the long tail already seems to have an essential aspect to its visualization, which is the basis for the name. Chris Anderson includes it as the logo on what seems to be the official blog about the long tail (can there be an official blog for a concept?). The visualization, in fact (or in “Wikipedian fact“), comes from a graph that follows a power law relationship.

There are probably a dozen ways that Amazon could implement this visualization into its page. The first way would be to simply use the vertical long tail in conjunction with a text link, which would appear similar to this graph of the Top 15 tags in the blog archives of Binary Bonsai (scroll down when you get there).

Another would be to use some AJAX-y goodness, which would allow them to use the more familiar, horizontal visualization, where the user could hover over different points of the graph and have an AJAX window pop up with the cover image, title, and new and used prices for the highlighted item. They already use AJAX throughout the site (hover over the “See All 32 Product Categories” tab on the Amazon home page, for example), and they even use the exact window I’m talking about in their beta-stage program for Amazon Associates, called “Product Previews.” As they tell their associates:

When users hover over a preview-enhanced link, a small window appears containing valuable content and information about the product you’re advertising, including an image, new and used price, average customer review, and availability. It also gives your visitors the ability to add the item directly to their Amazon.com shopping cart.

It couldn’t be that difficult to connect those AJAX windows to a long-tail graph on their own Web site.

Such a view into the “Customers Who Bought This Book Also Bought” long-tail variation would not just show me the range of books on the long tail, but it would also show me the “power relationship” within the dynamic. With the current basic-HTML list of links, I cannot be sure whether one person also bought the recommended book or whether a thousand did.

With such knowledge, I might be more persuaded to add a companion to my original item — after all, if almost everyone who buys James Joyce’s Ulysses also buys Don Gifford’s Ulysses Annotated, or even better, if I knew that more people buy James Joyce’s Ulysses, by Stuart Gilbert, rather than its seemingly direct competitor, Ulysses (Cliff Notes), by Edward A. Copper, then I, as the user, would know which of the competing books is more “trusted” by the market.

It might be even stronger if Amazon could, along with providing the long tail, inform us as to how many customers bought the various companion pieces when compared to the total number of customers who bought the original item. For example, if I know that 76% of the people who buy Milan Kundera’s The Unbearable Lightness of Being also bought Love In The Time of Cholera, by Gabriel Garcia Marquez, then I might think twice before deciding not to follow suit.

While it would be possible to “show” the long tail by simply including this percentage next to the text link, I think Amazon would be embracing the concept of the long tail more if it included the AJAX-ified graph. This would allow Amazon to utilize the entire long tail without having to make the page even longer than it is, or without having to make the quasi-autocratic decision as to what five books make the list of “Customers Also Bought…”

(I say “quasi-”, because if it shows only the top 5, then the customers are, in effect, making the decision; but even if this is true, there’s a reason why a democracy is scarier than a republic — I’d rather see both the minority and majority of purchases)

Of course, there’s the question of “where to cut off the long tail?” It probably wouldn’t be practical to link to every book that shares a customer with any given book, so Amazon would have to make some decision as to how many customers a given “shared relation” must have before it appears on the list. I’m not about to make a recommendation on that decision. But because the decision needs to be made is not enough reason not to pursue the feature.

I sure hope someone from Amazon reads this.