(Mis)Tracking Web Traffic

PreacherTom writes "Online advertising is considered by many to be the most dependably trackable ad medium of all time, with revenues expected to grow to $16 billion in this year alone. However, companies are finding that competing methods of measuring web traffic are giving contradictory results. Since advertising revenues are based directly on the traffic developed, this news could mean serious trouble. For example, valuations for startups such as Facebook and YouTube appear to be doubling every few months, but those numbers are based on traffic figures that could be misleading."

Comments

Customers 1/10th of IPs

[Anonymous Coward]

I told you a million times and I will tel you again. There is a fraction of real customers that corresponds to unique IPs. And cookies are not reliable. Check out my study, which is almost complete at this point. Results would be available next week.

Huh?

[Tarlus]

How is this "serious trouble"? Not to be a troll or anything, but why does it matter that much if web advertisers have inaccurate figures on their incoming traffic? Especially in a world that readily embraces advertisement and popup blockers?

So who is it, anyway?

[radarsat1]

So who the hell is actually clicking on all these online advertisements?
Who is actually responding to and buying things from spam emails?

I just don't understand how these kinds of things can be profitable, given that I've never met anyone dumb enough to fall for them. I certainly have almost NEVER seen an internet ad and said, "hey that's just what I'm looking for! CLICK."

I mean, I understand that it has value in the sense that it puts logos in front of peoples faces and reminds them about products and such, but where is the direct value in online advertising? No one honestly clicks on website ads on purpose.

Re:So who is it, anyway?

[HarvardAce]

I just don't understand how these kinds of things can be profitable

It can be quite unbelievable, and it applies to more than just online advertisements. It's hard to believe that a $2.5 million advertisement during the Super Bowl is profitable, but I trust that the companies that buy these spots have done their research and it is actually profitable. Overall Anheuser-Busch spent $850 million on advertisements in 2005. While not all of that is advertising beer (they have a few amusement parks for example), let's say that half of it was on beer and they make a profit of 50 cents per bottle/can or equivalent. That means their advertisements would have to increase sales by 850 million bottles of beer in order to make a profit. This is of course greatly simplified, but it is a good exercise to show how susceptible the population is to advertisements.

Re:So who is it, anyway?

[ljw1004]

"if a polite ad helps keep some site up and running, it costs me nothing."

So where does the ad money doesn't come, do you think? In 1999, spending in the US on advertising was about $6000 per capita. We ourselves are spending all this money, in the form of inflated product prices, to subject ourselves to the ads. How much of that money trickles into the websites we wish to support? I don't know, but I bet it's only a tiny fraction.

Ad impressions do not lie

[valtoid]

The most reliable form of web traffic measurement is reported ad impressions from a reputable 3rd party like DoubleClick. They do a fine job of eliminating bots and spiders (unlike log files) and only meaure an impression once an ad is served. Eliminate pop-ups and unders (which our sites do not take) and one can get a true measurement of traffic. If you have on average 2 ads per page on your site and monthly ad impressions served is 100,000, then you have 50,000 monthly page views. Sure you are still in effect undercounting when you factor in ad blocking, etc....but what do investors and VCs really care about, the ability to generate revenue. Ad impressions are the true way to factor that in and provide a fair measurement of traffic. ComScore and Neilsen are antiquated measurements of traffic, how many Slashdot users out there would download ComScore software to be a part of a web "panel"??? I don't think many of us would do that.

this takes me back

[xoundmind]

Remember the horrible old days when Webalizer stats were being misinterpreted wholesale? I inherited the primary coding responsibilities for a large consumer health site. It was about 10,000 static html pages and the site relied on Webalizer "hits" for years. Even reported them as such to the large pharmaceutical companies that sponsored the site. After we converted the site to being MySQL driven and actual page views were accurately recorded, the sponsors, editors and project managers FLIPPED. Like it it was the programmers fault that they had been so stupid all of those years....and perhaps acting fraudulently towards the advertisers. Mercifully, I worked on other angles than stats, but the other programmer caught hell for a year.
I'm sure others in the audience have similar horror stories.

Nothing new

[cronian]

Web advertising statistics have been manipulated and misleading since the beginning of the dot com era. While the internet advertising makes it easier to track usage statistics more accurately, site owners have strong incentive to lie. Furthermore, 1000 impressions != 1000 impressions.

There are lots of different tricks. If you selling by the impression, you can move the ad to a less visible spot on the page. You can also commit outright fraud, and just release the wrong numbers. As we've seen, clicks can be manipulated by bots. Internet advertising is more than anything about conversions and sales. How many people make it to your web page, and actually do something. If you're smart, you'll adjust your ad buys accordingly and ignore misleading statistics.