WordPress

How bad DO sites overstate their number of visitors?

EyeballsComScore just released its White Paper on Cookie Deletion. Cookies are little files that web pages access to save information in for marketing, analysis, analytics, and to assist with user experience. For instance, when you check a box to save your login information on a site, it’s typically saved in a Cookie and accessed the next time you open that page.

What is a unique visitor?

For analytical purposes, every time a web page sets a cookie, it’s marked as a new visitor. When you come back, they see that you’ve already been there. There are a couple distinct flaws with this approach:

  1. Users delete Cookies… a lot more than you think.
  2. The same user accesses a web site from multiple computers or browsers.

Regional news sites are able to charge advertisers based on information like this. In fact, the local Indianapolis Newspaper states, IndyStar.com is central Indiana’s No. 1 online resource for news and information, receiving more than 30 million page views, 2.4 million unique visitors and 4.7 million visits a month.”

So how much can cookie deletion skew numbers?

The results of the study reveal that approximately 31 percent of U.S. computer users clear their first-party cookies in a month (or have them cleared by automated software), with an average of 4.7 different cookies being observed for the same site within this user segment. Prior independent studies conducted by Belden Associates in 2004, by JupiterResearch in 2005 and by Nielsen/NetRatings in 2005 also concluded that cookies are deleted by at least 30 percent of Internet users in a month.

Using the comScore U.S. home sample as a base, an average of 2.5 distinct cookies was observed per computer for Yahoo! This finding indicates that, because of cookie deletion, a server-centric measurement system which uses cookies to measure the size of a site’s visitor base will typically overstate the true number of unique visitors by a factor of up to 2.5x, which is to say an overstatement of up to 150 percent. Similarly, the study found that an ad server system which uses cookies to track the reach and frequency of an online ad campaign will overstate reach by a factor of up to 2.6x and understate frequency to the same degree. The actual magnitude of the overstatement depends on the frequency of visitation to the site or exposure to the campaign.

Are advertisers being taken advantage of?

Maybe! Take a site like the local news site and that 2.4 million number instantly drops to under a million visitors. A news site is a site that is frequently visited as well, so that number could be well below that. Now add the number of readers that visit the site at home and at work and you’re dropping that number another significant amount.

This is trouble for the old ‘eyeballs’ crowd. While sales folks are always selling by the numbers, their websites could actually have far fewer visitors than competing media. Of course, there’s no real way to ‘fix’ the issue. Though any web professional with half a brain recognizes that this is the case, I’m not trying to state that sites are purposefully overstating their numbers. They aren’t overstating their statistics on purpose… they are simply reporting industry standard statistics. Statistics that happen to be very, very unreliable.

As with any good marketing program, focus on the results and not on the number of eyeballs! If you are comparing rates between media types, you may want to apply some quick math so the numbers are a little more realistic!

5 Comments

  1. 1

    Perhaps in the future something along the lines of CardSpace will illuminate this problem. Although, it could become too Big Brother. We’ll just have to wait and see.

  2. 2

    you said it, there is no accurate way of determining unique visitors to a website.

    cookies are not reliable and now many people are using flash for client side storage.

    But for advertisers, page view is all that matters. Its easy to accurately determine the numbers of times an ad is displayed 🙂

    And then, many web statistics services have there own problem. Live statistics site like statcounter will take into account a limited number of users at a time.

    google analytics is much better at this, but sometimes i have to wait for 2 days to get the latest report 🙁

  3. 3

    “an average of 2.5 distinct cookies was observed per computer for Yahoo!”

    How many Yahoo users are there per household computer? Yeah, probably around 2 or 3. I know I’m constantly logging off my wife so I can check my account, whether it’s on Yahoo or Google, Schwab or any other site.

    In our house, we happen to have 4 PCs and a Mac online between 2 adults, so it happens whether you have one computer or many.

    If you have a reg site and your server logs handy, make a report of the names for each IP addresses. (this shows how many people share computers/have dup accounts). Then make a report that shows how many IPs each name has appeared on. (this shows that a) ips get recycled by isps and b) users log in from mulitiple locations. )

    So yeah, the 2.5 number is about right. Not fraudulent, not overstated, just right. No story here. Move along now.

    • 4

      The article that is written is not discussing login/logout issues with respect to cookies, it’s talking about cookie deletion and its impact on unique pageviews. Yahoo! does not delete cookies when you logout and login.

      At issue is that over 30% of households DELETE their cookies, so you are seen as a new visitor… not another one in the household. Please read the article for a more in-depth explanation.

      Your example is also what I mention in my post, that many people visit the same site from multiple machines. With 4 PCs and a Mac between 2 adults, if you visit the same site on all machines, you can be seen as up to 5 ‘unique visitors’, not 2.5! And if you are regularly deleting cookies as do 30%+ of the population, that turns to well over 12.5 unique visitors.

      Like I said, I don’t believe it’s fraudulent… but it IS overstated. Your household proves it.

      Thanks for commenting!

  4. 5

    Re-reading the article and your response again…you’re right. I originally misunderstood your point. Thanks for clarifying.

    That being said, gautam is right–more and more folks are using flash cookies, even when they have no other reason to serve flash. Dirty little secret: you can’t (easily) delete cookies set in your flash.

    (Google’s doesn’t serve much flash. DoubleClick does…)

    If sites want to come clean to advertisers, they need more transparency about how many times which object was viewed by whom, and when.

    Since log files aren’t good at that, they”ll need lots of data in a database. A very large database.

    Since that isn’t going to happen soon, the best idea, like you say, is to focus on results!

Leave a Reply