Heck Of A Guy

"Most Viewed" Means "Most Viewed" – Mostly

One of the features recently added to the Heck Of A Guy Blog, the “10 Most Viewed Posts” list,¹ has drawn a number of email queries about the means by which  the ranking is  calculated. These inquiries, in turn, led to my plan to  publish a brief and simple explanation of how the listing is generated and what it does and does not mean.

Well, that was the plan. To paraphrase Tina Turner, only there’s just one thing, I never write anything brief and simple. I always write it long and complex.

In this case, it turns out that the  issues that can cause confusion about this simple list of most viewed posts are also pertinent to interpreting election polls, medical research, and reported statistics in general.

Consequently, today’s post will focus on an overview of the mechanics of this small area of web statistics and the lessons it may hold for understanding references to research and polls in the popular press, daily conversation, and the professional literature. A follow-up entry will look at the actual posts on the most viewed list and how they got there.

Counting Views With WP-PostViews

The referenced list on Heck Of A Guy is automatically generated by a WordPress Plugin called WP-PostViews, which, in the words of its author, “enables you to display how many times a post/page had been viewed.” In the support section, he explains, “Technically WP-PostViews will count every time a hit on any posts/pages even if it is just a refresh as it does not log any data.”

I admire the simplicity – count the hits, show the number of hits for each post, display the posts in rank order, and that, as they say, is that.

Except, it’s not exactly that simple. First of all, the user can choose to prevent one or more specific posts from appearing on the list. I hasten to add that this is a legitimate option if used appropriately. One blogger wished, for example, to omit the page with the most hits on his site because it contained only contact information. It seems within the spirit of the “Most Viewed” list to eliminate from contention those posts and pages that include only housekeeping data. In another context, if one were interested in the presentations at a medical conference that were best attended, one might choose not to list a meeting that was attended by everyone but was concerned exclusively with registration, schedules, credentials and the like if the other sessions dealt with new medications, clinical studies, morbidity and morality data, etc.

The blogger also has the choice of counting only registered users, only guests (aka normal surfers), or both.

Again, both of these options are legit functions, but even when used benignly, they clearly change the results.

More confounding still are parameters of happenstance. For example, I installed this plugin within the past month or two, after Heck Of A Guy had been running for more than two years. The rankings are established by comparing the cumulative hits of each post, Consequently, in this intra-blog competition for most hits, older posts receive no credit for hits that took place before the plugin was implemented. Less obviously, it puts older posts at an additional disadvantages. There is a group of readers who view every Heck Of A Guy Post within a day or two of publication.² Those entries that went online in the pre-WP-PostViews era forfeit those automatic viewings. Posts tied to contemporary issues enjoy a plethora of hits while contemporary is contemporary; when contemporary is two years old,³ the hits typically diminish dramatically.

Most importantly, the number of hits provides no indication of the motivation for the viewing. A reading of this post by a University of Chicago graduate student as research for his never to be competed doctoral dissertation, “The Hegelian Dialectic Manifested In Optional Secondary Purpose Software Choices Made By Midwestern Bloggers In The Early 21st Century,” a visit to yesterday’s post by a bot gathering data for a search engine, and the 28th time your three year old granddaughter punched refresh to watch the video of the ducklings jumping from their nest⁴ each count as one hit.

Alternative Approaches To Determine Post Popularity

Other chunks of code are available that perform a function similar to that provided by WP-PostViews. There are several more plugins for WordPress alone⁵  that provide a list of the “Most Popular Posts.”

Most Popular Posts equates popularity with number of comments, i.e., the post with the most comments is #1 on the popularity list, the post with the second largest number of comments is #2, and so on.

Some, like WP-Plugin MintPopularPostsWP, require the blogger to sign up for a specific stats service or use an independent traffic tracking tool, which then generate results which the plugin displays.

One of the best-known WordPress plugins that develops and publishes a list of  most popular posts is Popularity Contest. Its mechanism is described at the plugin’s web site this way:

Popularity Contest keeps a count of your post, category and archive views, comments, trackbacks, etc. and uses them to determine which of your posts are most popular. There are numerical values assigned to each type of view and feedback; these are used to create a ‘popularity score’ for each post.

The values assigned to each view and feedback type are editable and can be changed at any time. When you change any of these values, the ‘popularity score’ for all posts are updated immediately to reflect the new values.

While still other applications designed for this same purpose exist, those already considered establish the primary point – there are numerous means by which one can create a most popular or most viewed or most requested post. The mechanical means chosen – and often the specific settings of the software chosen – will change the criteria for and thus the meaning of “most popular.”

Most Viewed Vs Most Popular Vs Most Important Vs …

The variations in the precise interpretations of the Heck Of A Guy blog’s 10 Most Viewed Posts list may arguably have less impact on the course of world events than, say, this country’s November elections, but I would hold that a consideration of the former feature may be helpful in understanding the latter event.

As discussed above, even something as seemingly straightforward as “most popular post” can be, intentionally or unintentionally, ambiguous or misleading. Depending on the software and how one adjusts the software’s settings, “most popular” is defined by a formula that may consist of any one criterion or any weighted combination of parameters. “Most popular” may be determined by the preferences of everyone visiting any page of the site at any time or, at the other end of the spectrum, a specific subgroup(s) of viewers that click on selected portions of the site during a specific time frame or the factors that exist at any point between those two points.

Moreover, it is the rare site that will include an explanation of the statistical calculations. To be fair, I do believe that the absence of those explanations is most frequently motivated by non-malignant causes, including blogger fatigue, aesthetics, the trivial nature of an intra-site post popularity ranking, and the conviction that no one would read a detailed report of the statistical processes chosen. I am less convinced that those same reasons pertain to polling reports promoted by a candidate for political office as proof that the majority of constituents support that candidate’s position on an issue or to the results of a study published in a news story proclaiming that a cure for the common cold has been discovered, the end of the world will be the result of terminal ennui caused by movies made by Al Gore, or housing prices will always go up.

The issues implicit in that sardonic observation will be addressed in a forthcoming  post in this series, tentatively titled,

The Official Unauthorized Heck Of A Guy Practical Guide To Interpreting Polls, Results Of Studies, Web Statistics, Performance Evaluations, Restaurant Ratings, Goat Entrails, Ratings Of Restaurants Serving Goat Entrails, Sociological Trends, Trendy Sociologists, And Reports Of The World’s Fastest Growing Sport.

Until then, I leave you with an example of this sort of problem, which returning readers may recognize as a set of excerpts from a February 2007 post,  Lessons Learned Willy-Nilly From Milli Vanilli.

An Instance Of Meaningless Popularity

The most popular real estate on the Heck Of A Guy site is
[pause for effect, 2, 3, 4 ...]
a 350 X 252 patch of pixels bearing the image of Milli Vanilli

That’s right. The prestigious Pixelated Post Popularity Prize goes to a publicity photo I employed solely to illustrate an example of pop musicians I disliked in an article otherwise devoted to artists who had won my adulation.

I Can’t Believe It’s Not Music [the post containing the Milli Vanilli graphic] did OK with readers, scoring in the top 25% of 2006 Heck Of A Guy posts by pageviews and unique visitors.

The Milli Vanilli photo, on the other hand, … well, that other hand has apparently been busy clicking away full time on that photo’s link in Google Image Search, accounting for thousands and thousands of hits, almost all of which are from different addresses.

I, for one, admit to being surprised to find that many people searching for a picture to represent musicians they don’t like.

But wait; there is – as you may have guessed – more.

The Perverse Perseverance Of The Milli Vanilli Graphic

The Milli Vanilli photo continues to be the most popular chunk of Heck Of A Guy geography as of this date [February 2007], which wouldn’t be surprising and might even be expected – except for the following:

The Milli Vanilli photo continues to be the most popular chunk of Heck Of A Guy geography as of this date – even though two months ago I wrote a robots.txt⁶ to prevent search engines from indexing the photo.

In fact, the Milli Vanilli photo continues to be the most popular chunk of Heck Of A Guy geography as of this date – even though I deleted the picture from the site altogether over a month ago.

The explanation for this perpetual popularity can be expressed as the Google-altered version of a standard proverb:

On Google, A Picture Not Only Is Worth A Thousand Words
But Also Lasts A Thousand Times Longer

While Google and the other search engines re-index the Heck Of A Guy text every day or two and important, fast changing sites on what seems an hourly basis, images are surveyed by a separate bot and re-indexing of graphics is accomplished at a more leisurely pace, perhaps as infrequently as every six months, according to some educated guesses.

But why did I dump the most popular file on the Heck Of A Guy site? That brings us to Milli Vanilli Photo Lesson #1:

Mom Was Right: You Can Be Popular For The Wrong Reason

If I had been selling a product that was in the picture or promoting the subject of the photo, I would, of course, been happy about its multitude of visitors.

If I were selling ad space, I might have kept the photo in place to promote the site’s popularity in general.

This was not such a case. The Milli Vanilli image was only incidental to the Heck Of A Guy blog. The image-hunters were not interested in the post that surrounded the photo, nor was the Milli Vanilli shot an original photo I snapped. They didn’t love me; they only wanted my Milli Vanilli.

Consequently, I pulled the plug, bidding farewell (I thought at the time) to all those hits. And, that segues into the most significant revelation of this series, AKA Milli Vanilli Photo Lesson #2:

Between Attracting & Pandering To An Audience Is A Thin Line

I am chagrined about the intense ambivalence surrounding the decision to delete that graphic. Even though I knew that the traffic it generated, as noted previously, had almost nothing to do with my blog and, if anything, it was a minor hassle because it used bandwidth from my account and confused the stats, I hesitated before pulling the trigger.

My own personality quirks aside, I am struck by the attraction I felt from this ephemeral, diaphanous flicker of esteem from folks registering approval of the blog by visiting the Heck Of A Guy site, regardless of their reason.

For the first time, I experienced an empathic understanding of the artistic dilemma novelists, poets, performers, and other creative sorts face when they must choose between adhering to their own private vision, which may restrict the potential audience and forgoing such purity in hopes of making oneself more accessible and attractive to a wider constituency.

After my truly trivial Milli Vanilli moment, it is difficult for me to grasp the enormity that allure must take on when creative work is the focus of ones life, not to mention the source of ones livelihood.

There you have it. “Most Popular” was, in this case, the result of the particular processes of Google Image Search, the unexpected and inexplicable perpetual popularity of Milli Vanilli, and my arbitrary choice of an example of singers I don’t enjoy and the even more arbitrary choice of a photo of that group.

That, my friend, is what you call “meaningless popularity.”

Those feeling brave may wish to consider such a phenomenon in the context of, say, the polls in the recent presidential primaries.

Scary, eh?

_____________________

1. See left sidebar, second set of links from the top [↩]
2. It turns out, by the way, that this group comprises individuals who are intellectually superior, extraordinarily perceptive, hard working, and easy on the eyes. [↩]
3. E.g., Gay Games Rowing Event Brouhaha In Crystal Lake [↩]
4. See Ducklings Provide Message Of Courage [↩]
5. I assume other blogging platforms have analogous software, but I’m familiar only with WordPress [↩]
6. A “robots.txt” is a script that directs search engine bots to index or not index certain files or directories on the web site [↩]