"You may fool all the people some of the time, you can even fool some of the people all of the time, but you cannot fool all of the people all the time.  P.S. Or webmasters about inflated site stats." - Abraham Lincoln

Ol' Abe.  What foresight!
 

First off, let's understand that in most cases of inflated stats we're talking about sheer, raw, unadulterated lust for money.  Like the Nielsen Ratings affect how much a TV station can charge its advertisers, and the Arbitron Ratings affect radio, so it is with Web "hits" and the sites that measure them.  The more hits, the more the higher-paying advertisers will flock to you.  So this is serious business.

There are a number of ways to inflate one's hits; some semi-illegitimately (through ignorance or bad programming), but most are a case of outright fraud.  Although I'll use specific sites as examples here, I don't mean to single them out as this is a net-wide problem.

Note 1:  There are actually three ways to determine site traffic; "hits", "visits" and "unique visitors".  An expanded definition is here.  For the sake of readability, I'll just call one visit to a page by anyone a "hit".

Note 2:  While you may know that you can jump over pages when backing up in the browser (by using the History box), most people don't.  Normal people, for the most part, really don't understand what's taking place when they're on a Web page or how to use a browser properly, and will just sit there hitting the 'Back' button until their eyes glaze over — with each click acting as another page view or 'hit'.  As such, I'm taking the whole "Back button" thing seriously.
 

Off-Linking 

You'll notice that most sites open a new browser window or tab when you click on a link that goes to another site.  This is called "off-linking", as referred to 'normal' linking to another page on the same site.  You'll notice that on The Drudge Report and Instapundit, two sites with extremely high hit rates, the linked post stays in the same window.  This way, when you're finished with the article, you have to hit the 'Back' button on the browser to bring you back to the original site.  That counts as a second hit.  Read three or four articles, hit the 'Back' button every time, and now the original site has you down for four or five hits, whereas you really only arrived at the site once.
 

Auto-Refresh 

This method is far more insidious and can really jacks up the stats.  If you look at the source code of Drudge's home page, you'll see this: 

var timer = setInterval("autoRefresh()", 1000 * 60 * 3)

That automatically refreshes the page every 3 minutes.  (1000 x 60 x 3 = 180,000 milliseconds divided by 1,000 = 180 seconds = 3 minutes)  And it's easy to prove.  Just go to his site and start a timer.  Somewhere around the 3-minute mark the screen will flash briefly as it refreshes.

So, as long as you've got Drudge open, no matter what you're doing on the computer (or if you're even at the computer), you're acting as a fresh hit every 3 minutes.  If a mere 1,000 people do this for an hour, that's a whopping 20,000 hits without anyone even clicking a link, so the numbers can really add up. 

And this is from Instapundit's home page: 

META HTTP-EQUIV="refresh" content="1800"

That refreshes the page every 1,800 seconds, or 30 minutes.  No big deal, you say? 

What makes this particularly insidious is the advent of tabbed browsers, like Firefox and IE 7.  This way, people have link sites like Drudge and Instapundit constantly open in one tab while they're off reading things in other tabs.  And what's going on the whole time that link site is open in the first tab?  Refresh, is what's going on, and your helpful — if unknowing and inadvertent — contribution to the site's ad revenues. 

So, if 10,000 news junkies show up at work in the morning, fire up Drudge on a tab and leave it up all day long (some of which time they may actually spend working), that's:

10,000 hits x 20 refreshes/hr x 8 hours = 1,600,000 hits/day

Now add to that all of the people who don't know the page is automatically being refreshed and are hitting the browser's 'Refresh' button to see if anything new has been added every time they return to it.  Scary, huh?  A mere 10,000 visitors could turn into two million by the end of the day, just from the page sitting there in a tab on somebody's work computer.

The site owners would argue that they're merely refreshing the page for the users' benefit, so that they won't miss the latest hot post, but that argument falls apart two ways:

• Remember, when you click on a Drudge or Instapundit link, you're staying in the same browser window, so their pages automatically refresh when you go back to their site.

   
• Browsers in their current form have been around for almost two decades.  Everyone knows where the 'Refresh' button is if they want to refresh the page to catch the latest, hot, can't-miss links — and that's especially true with the type of people who are into the latest, hot, can't-miss links.

 

And, right on cue as I'm putting this article together, there's ol' Matt with the staggering February numbers of 568 million hits.  Remember, because he doesn't off-link correctly, if you assume every person who went to his site clicked on just one link, then you immediately have to cut the total number in half, due to having to 'Back' your way back to the site, thereby creating another hit.  If the average person clicks on two links, then you have to cut it by two-thirds.

Then you add tabbed browsers and the whole "Auto-Refresh every 3 minutes" thing into the equation and you have to wonder what fraction of that 568 million the number really is.
 

The Redirect Code

I'm sure you've noticed that occasionally you'll go to a site and the 'Back' button doesn't work.  This is because the clever webmaster has put a 'redirect' code in the source file, so every time you hit 'Back', it redirects you back to the current page.  The primary purpose of this — I have to presume — is that you'll eventually give up and take a look around, possibly buying something or at least bookmarking the site — so that later you'll buy something.  But the secondary benefit is certainly that those ol' hits are mounting up as people click in vain to escape the site's clutches.

As I said earlier, most people really don't understand what's going on, computerwise, and that's especially true with older people.  They'll either think they're doing something wrong as they continue to click on the 'Back' button, or they'll think the Internet is suddenly down and if they click on it long enough, it'll eventually work.  I taught at a computer college for two years and did field work and gave seminars and tutored people for a number of years and saw this kind of stuff all the time.
 

Coding Chicanery

Here's the 'History' box in Explorer 6 after going to Fox News, reading a number of articles, including articles linked from articles, then backing up to the home page: 

That's exactly as it should look (on my system).  "Start Page" is the links page on my computer that the browser opens up with.  Fumble around all day long inside the average site, back your way up to the home page and you should see one entry, like the above, in the History window. 

Unless, of course, you're MSNBC: 

So, after doing the same routine I did with Fox, the average slob ends up hitting the 'Back' button five times just to get back to where he started.  And each one of those is another hit.  "My, that MSNBC site sure is popular!  Just look at their stats!"

What's puzzling is that it doesn't act this way all the time.  A week ago, when I was writing the first draft of this article, it suddenly was operating correctly, whereas usually it acts like the above.  A cynic would say that on some particular days of the month some governing body checks out the integrity of certain professional sites and that's when the webmaster cranks it back to normal.

But I would never suggest such a thing, of course.  I'm sure it's just a case of...
 

Coding Errors 

I'm going to presume this is just bad programming, rather than something a little more nefarious. 

There's some blogging software around with the somewhat oxymoronic name of "Serendipity" (whereas blogsites are about the least serendipitous spots on the web) that has a small bug that, over time, could inflate the hits a tad.

Let's take a stroll over to a site that uses the software.  Hey, there's an article on Hillary that I want to respond to!  I furiously scribble down my little missive and hit the 'Preview' button to see how it'll look.  Realizing that my comment might perhaps be read by millions and change the very course of civilization, I carefully comb through it, adding this, amending that, all the while hitting the 'Preview' button to see how it'll look.  Finally, satisfied that my important comment is worded correctly and is ready to alter the tide of history, I hit the 'Submit' button and off it goes.

Okay!  Well, let's hit the ol' 'Back' button and see what else is on the home page...

I mean, let's eventually see what's on the home page.  Due to a bit of poor programming on Serendipity's part, the browser is actually logging each 'preview' as a separate page in the history and it'll take me seven clicks on the ol' 'Back' button to get back to the home page.  So, although I really only went to one page on the site, it'll end up logging about fourteen hits total, one for each of the six 'previews', the six to get back to the home page, and two for the home page, itself.
 

The Ad Angle

Normally, on a site that doesn't carry advertising, whether the stats are inflated or not really doesn't matter.  The problem arises when the stats get to the point where it begs the question, "Should we start advertising?"  That's a pretty big jump, in the sense of setup time & cost to get some advertisers online, not to mention getting a business license, dealing with the IRS, etc.

But, as it's always been on the Web, it's all about the numbers:

If you get 1 million monthly visits to your site and just 1% of them click on a banner that makes you one thin, crummy little dime, that's a thousand bucks a month of pure gravy.  So you can see why it's tempting.

One problem is that stats can be grossly overinflated and the owner might not even know, like if the original webmaster slipped an AutoRefresh into the source code while constructing the site, thus giving the owner the illusion that advertising is a viable option.  Another problem is that you're bound to lose a certain core readership if suddenly the site 'goes commercial'.

And there are further problems, such as the 'moral dilemma' of what ads to run.  Do you run the big "IMPEACH BUSH!" banner (5,000 hits/day) or the one for this really excellent book you just finished (3 hits/day)?  Do you run the crass "SHOOT THE DANCING MONKEY!" banner (1,500 hits/day) and lower the overall IQ of your site by 50 points, or do you run the respectable and stately Verizon banner (7 hits/day)?

And then there's the biggest problem of all.  Once "numbers" are the game, you start thinking about altering your format or 'site theme' to draw more people.  You start writing articles for their linkability and popularity, rather than from the heart.  You bring in co-bloggers to sparkle the place up, thereby driving up the numbers and, in the process, destroying that undefinable element that made the blog special in the first place.
 

Hit Counters

About the last thing you should believe on a web site is a hit counter.  I did an AutoRefresh experiment years ago using a counter and the numbers were spinning around like a slot machine.  And you can start the numbers anywhere, just to note.  I've, uh, heard that some unethical webmasters start their clients' counters at "10,000", just to give the new business an established look.  Hey, if a bank can start out a checking account at check #1000, then surely I those unethical webmasters can do the same thing!
 

'Raw' Hits

When it comes to 'raw' hits; that is, without any further information such as which pages were accessed, there are a bunch of ways to generate hits.  You could have a secret page that has an AutoRefresh set to 1 second, so basically it reloads itself the second it's loaded, then open a hundred of them on your own computer and just let those ol' hits mount up.  Ditto running a DOS batch file and ftp'ing the site every 2 seconds (times 100 batch files) 24 hours a day.  It wouldn't count as far as advertisers and those who know how to dig up the stats, but it would sure impress the newcomers.

This site has received   130,563,642 HITS   in the past day!

 
Summation

Unfortunately, it doesn't appear that the fraudulence out there is going to be cleaned up anytime soon.  There are groups pushing for this or that measuring standard, but there will always be ways to pad the stats.  If not internally with tricks like AutoRefresh, then externally with actual computers faking their way in using randomly-generated IP addresses and the like.

Bottom line?  Don't believe a single stat you see.