how does this affect advertisers?

for a marketing blog that sizzles, check out

marketmpb.blogspot.com

matt

Seems a bit strong, making public criticisms because you think a vendor made an error of 3%. That margin of error in an industry that is often questioned for its inaccuracy is pretty good.

Comscore and all this scoring websites are pretty much way off or is basically a liar. How can they measure a website’s hit without installing a script in the webpage. Comscore and others overreport by at least 20% to 50% or underreport. It is just a money making scheme.

Err… Except they do have scripts for measurement: direct.comscore.com. Also
yahoo claims they are 3% off… Where’s the 20% coming from?

I noticed that a website I know that is not in anyway associated with compete (maybe not comscore) had their traffic counted by Compete and the traffic numbers were way off. Also I doubt Yahoo installed comscore traffic script on every single page. Comscore then suddenly produces a report saying we counted Yahoo!’s traffic is just ludicrous. Put the script on the webpage and legitimately count it, or don’t count other people’s website without the script. It just like producing a story. It is just a publicity and money-making, catchword scheme.

Lol. It is coming from an experience from Compete (different company) that reported a website that I know. It overstated the traffic hit by almost 200%-300% percent from the actual level. Also the numbers they produce are generalizations and is always on the 20 million to 30 million range. So that is a lot of rounding. It is also possible that they can undercount by the same percentage. That is my guess. I’m saying these things are very wide generalization.

I work in web analytics- thought I’d throw in some info about ComScore and traffic reporting in general:

1) There are many, many ways of measuring traffic on a website, from ComScore to Quantcast to Google Analytics to Omniture. ComScore is important because it’s one of the main tools potential advertisers look at when gauging audience size. The reason for this is that ComScore’s measurement is third-party, so site owners can’t really manipulate the data to make themselves look better.

2) ComScore measures traffic in two ways- estimates based on statistically significant samples of US and worldwide visitor, and by direct measurement through installation of a tracking beacon. Most major sites have installed ComScore’s direct measurement beacon since it provides more accurate reporting than the panel-based measurement. That said, ComScore numbers aren’t always going to match Google Analytics’ numbers. Advertisers and site owners look at ComScore’s numbers directionally.

2) A 3% error is a big concern for a site as large as Yahoo!. ComScore was basically saying that they had 3% less inventory available during June, or about a billion fewer page views. Even at a 50 cent CPM (which would be the low end of what an advertiser would pay on Yahoo!) and just one impression per page view, that’s more than half a million dollars in potential revenue.

The difference is that Compete is entirely panel-based measurement (and notoriously inaccurate) and ComScore is the IAB standard with hybrid panel and direct measurement.

You’re comparing two entirely different things.

“Where’s the 20% coming from?”

Out of his ass… it seems people want to make their opinions and make them facts.

From Comcast’s statement: “…. a processing error confined to data specific to Yahoo! …. ”

This is why Yahoo is pissed. Comcast was twiddling Yahoo’s numbers specifically and made a mistake. Why special twiddling for Yahoo? No one knows and Comcast isn’t talking.

Comcast denies all involvement.

So what’s the point. Y! will still be a loser company. http://2su.de/bMS

Completely agree. Every provider uses different algorithms to measure web stats, so that even direct measurement tools such as GA, Quantcast and server-based metrics will report different results. 3% is well within the margin of error and it seems quite petty for Yahoo! to be making a stink about this.

At the end of the day, the 3% shouldn’t make a difference in Yahoo!’s revenues since they will be paid on impressions delivered, not estimated.

In the online advertising world advertisers spend their money where the eyeballs are at. So for Yahoo to be undercut by nearly 1 billion page views, that could translate into missed revenue opportunities.

Advertisers follow these traffic reports very closely to help them understand where they should allocate their ad spend. For example, a lot of advertisers are starting to spend money on Facebook because of the amount of visitor traffic (eyeballs) they are receiving.

If I’m an advertiser and I have a choice to put ads on a site that gets 2 billion page views vs. a site that gets 1 billion page views…guess where I’m spending my ad budget?

Wouldn’t you be pissed off if you owned the site with the reported 1 billion page views when in reality you actually had 2 nillion page views?

1 billion, no matter what % of total, is still a big freakin’ number!

Josh, think of this as the expected margin of error (whatever % that would be) that advertisers expect from ComScore and would affect all sites equally due to the reasons you mentioned PLUS an additional 3% due to a processing error that only affected Yahoo.

Yahoo makes money on there marketing and page hits being off a billion is a big issue. When it comes to sales and ad views accurate numbers are very important.

Most people in the online media and advertising industry know that comScore’s methodology is inherently flawed and inaccurate (panel vs census – still not sure why they’d use a panel in the most measurable medium ever…). In particular, smaller properties exerience even worse under-reporting than larger properties. comScore has two things going for them: a) there’s nothing else yet widely accepted; b) their inaccurate measurements are fairly consistent across the board.

COMSCORE’s data is not reliable.
COMPETE’s data is not reliable.

Use them for directional (trend) purposes only.

If you ever see a Comscore number that purports to represent market share or actual #s of visits or sessions, just recognize that it could be off by a factor of 50 to 100%.

If the site gets fewer than 10 million uniques a month, then that factor may be more like 100 to 300%.

You can take those facts to the bank.

It’s tough to really know how much traffic Yahoo is getting unless you look at all tracking companies combined as well as internal data. Either way if they start adding in their content mills like with Associated Content I think it will knock the CPM’s of these pages down significantly. Advertisers today are looking for quality traffic to their websites not someone who accidentally clicks their ad.

I’m still waiting for comScore to go out of business.

Might take some time… http://yhoo.it/bekmF3

Hmm … their Page Views for Germany are still 26 percent lower than what is measured in Germany by the “official” measurement system for June even after the correction. Looks like their measurement Kung-Fu is pretty weak.