Lies, Damned Lies & Statistics: ComScore's Facebook Audience Claims Are Wrong By A Factor Of 17,000!

Audience measurement may be an imprecise art, but the figures that the top research companies provide to business can swing the expenditures of many billions of dollars. Television ratings, for example, are responsible not only for the survival or cancellation of every show on the air but also the determination of how much a station or network can charge for its commercial time. It is therefore imperative that audience measurement be as precise as possible.

When ComScore, one of the leading online research firms, issued a report on how much time web surfers spend on Facebook & Google, the figures hit me between the eyes like a freight train. They weren't just wrong... they were impossibly wrong.

The ComScore Statistics Aren't Just Wrong... They're Impossible

ComScore Inc. is a leading internet marketing research service company with a long standing reputation for in-depth statistical analysis of online behaviors. Countless companies base much of their online activities based on the reports generated by the company, which employs over 700 people from its base in Reston, VA. The company and its executives have won several prestigious industry awards. It is ranked among the Global Top 25 market research firms and the World Economic Forum named it one of 47 innovative companies. Its current market cap is close to $600 million.

 Like this article? Please Digg It!

With such an exalted status, it would seem almost impossible that they would be fabricating figures out of thin air, but it seems that is exactly what ComScore Inc is doing.

The company recently made worldwide headlines by releasing a report that American web navigators are spending more time on Facebook than on all of Google's sites put together. The ComScore report stated that in August, 2010 41.1 million minutes were spent on Facebook, while Google was able to capture only 39.8 million minutes of web surfers' attention. This report was carried on all major media, including CNN.

ComScore's Claims Are 17,000 Times Short Of Reality

Seems very impressive at first... until we pull out the old calculator and crunch a few figures.

OK... 41.1 million minutes... That's 685,000 hours. No matter how you crunch it, that's a lot of hours... but let's dig a bit deeper. Facebook, on their statistics page, claims "People spend over 700 billion minutes per month on Facebook"

Yikes. That's a bit of a difference. We're not talking about the difference between 41 million and 700 million which would be a time claim 17 times smaller than that made by Facebook itself, but a claim that it's 17,000 times smaller!

The Equivalent Of Only 5,800 Americans Watching The SuperBowl

Let's place this in perspective. If the same sort of discrepancy had creeped into top ComScore competitor AC Nielsen's television ratings, then instead of 98.7 million people reported watching the last SuperBowl, the total would have been 5,800. It doesn't take a PhD in statistics to figure out that there are some American neighborhoods where more than 5,800 people watched the SuperBowl, let alone in the whole country!

If ComScore Is Right, Each Facebook User Is Online Half A Second Per Day

Could it be a case of Facebook exaggerating? Unlikely as 685,000 hours spread among Facebook's 167 million US users would mean that each one is accessing Facebook for 14.8 seconds per month, or less than half a second per day. That's not exactly enough time to even load a single page, let alone plant an entire field full of crops in their irrationally popular Farmville game which just by itself claims 100 million users.

ComScore Caught In A Stupid Lie That Wouldn't Fool Anyone With A 99 Cent Calculator

So what happened? ComScore lied. A bald-faced, ridiculous, stupid lie which would have been caught by any high school dropout with a 99 cent calculator. They fabricated the statistics to get publicity, and they did so in such a ridiculously amateurish fashion that they have now effectively demolished the elevated status and credibility of the entire $600 million company. After this debacle, they shouldn't be trusted to take attendance at the local elementary school.

This is not ComScore's first rodeo. They have been linked on at least two occasions with allegations of distributing spyware and rerouting user traffic through proxy servers. Not exactly confidence-inspiring from a Global Top 25 research firm!

High-Priced Journalists Parroting Press Releases Are A Disgrace

Don't any of these high priced journalists mindlessly reporting this dreck know how to use a calculator? Or do they just parrot any press release that lands on their desks? Is it possible that they misinterpreted the figures? That's not how it seems on the major media reports I've been looking at. They're very clear: "U.S. Web Users spend..."

Which one is worse... if ComScore lied, or if the major media doesn't know the difference between millions and billions?

Heads Have To Roll: Either At ComScore Or At Major Media

ComScore: If it's proven the figures come from you, admit that you were wrong by a factor of close to 17,000 or justify your figures now. Don't try to spin out of these moronic lies! You're caught red handed, now retract immediately and justify how your entire 700 employees could be so stupid! Either that or if it is the journalists who can't read, then the major media should be handing out a whole whack of termination slips to their reporters tomorrow morning.

Heads have to roll on this!

CNN
CNN
Computerworld
Computerworld
ABC News
ABC News
CBS News
CBS News
DailyTech
DailyTech
AFP
AFP
International Business Times
International Business Times

Later development: Facebook's claims of 700 billion minutes seem to be worldwide, so if we cut it down by 30% which is the stated number of US users we're still looking at 210 billion minutes vs. 41 million: over 5,100 times. Under no conceivable formula can the 41 million for all US users be anywhere near correct by several orders of magnitude.

More by this Author


Comments 14 comments

6 years ago

Yes, the numbers are absurd, but... have you seen ComScores' original statistics and methods? I haven't, they are probably for pay. I can't be arsed to search really. I'm speculating here.

But it could very well be that something got "lost in translation" there, either in ComScores' reports, or the media's (considering the average quality of journalists regarding "technical" things like basic statistics' concepts and science, it wouldn't be surprising). I don't know how ComScores monitors these things, but if it is like the old time TV audience stats, they would be monitoring a certain amount of (hopefully) representative users instead of the whole population, and counting the usage patterns of *those people only*. You know, the type of thing statisticians do -- because if you observe the WHOLE population it's not statistics, it's a census.

So, maybe the 40 or so million minutes where spent by the sample monitored, NOT by all the possible users summed, and someone screwed the pooch in the reporting of the whole thing. The significant numbers then would be the *percentage* of time spent in which activity, which should be the same as for the whole set of users represented by that sample (if it was selected appropriately).

It would be interesting to know what the truth actually is in this case. Just for laughs.


Hal Licino profile image

Hal Licino 6 years ago from Toronto Author

Thanks for your comment. I'm also suspecting that the figures might be from the sample group, but I haven't seen anything that might substantiate this. In that case it would definitely beg the question of how all this media reporting MISSED that.


frymaster 6 years ago

one other issue is that facebook is reporting global user stats whereas comscore is reporting for the US only. while I don't know if facebook is popular in china, it's definately popular in english speaking countries and western europe, and, cruicially, india. That being said, I doubt the global number of facebook users is 17,000 times more than the number of US users, but it's a start. You also then have to go into how you calculate "minutes spent on a website" which is technically difficult:

- page views don't tell you how long a person's been on a page; a half-hour gap between page views doesn't mean the user spent half an hour reading the first page, it could mean they are queueing up tabs to read, and you've no way of telling how long someone spent on the last page of their visit

- facebook can use the continual background ajax queries to solve the second of those 2 issues, but then you have the issue of people who just leave facebook open all day and check it every now and again, also the issue of people who use multiple devices (like, say, phones)

so, the figures are wierd, but i'm not immediately convinced they're wierd because of fraud or even incompetance; I'd like to see both facebook AND comscore's methodolgy first


Hal Licino profile image

Hal Licino 6 years ago from Toronto Author

This Hub article clearly deals with USA only and the figures are completely impossible. The number of Facebook users in the USA is 30% of its worldwide total, as per FB's stats page which is backed up by their server records so I'm pretty sure it's spot on. FB's claims of 700 billion minutes SEEMS to be worldwide, so if we cut it down by 30% we're still looking at 210 billion minutes vs. 41 million: over 5,100 times. There is noooooooo way that the 41 million figure can possibly be correct for all USA users. It has to be ComScore's target group and misinterpreted by the media, or a whopper on the part of ComScore that is truly epic.


quicksand profile image

quicksand 6 years ago

There is an e-mail marketing company that undertakes to blast your mail once in seven days to nine billion people! Three billion more than the entire population on this planet! :)

PS

They use figures, 9,000,000,000. Probably they thought a million had nine zeros instead of six. That is the only explanation.


Hal Licino profile image

Hal Licino 6 years ago from Toronto Author

Actually, I know that company. Their claims are correct, as they also send to Vulcans and Klingons. :)


quicksand profile image

quicksand 6 years ago

My apologies to that company and to you too! Cheers! :)


Hal Licino profile image

Hal Licino 6 years ago from Toronto Author

I look forward to receiving my ngoqDe' (message or email in Klingon) whenever I'm patrolling the Neutral Zone to keep an eye on the Romulan marauders!


quicksand profile image

quicksand 6 years ago

Just a word of caution, I hear Romulans unlike Martians, Vulcans and Xonufilozhicans eat metal. Your motor cycle is in danger of being chewed up! :)


Hal Licino profile image

Hal Licino 6 years ago from Toronto Author

Actually I have to correct you about the damn Xonufilozhicans. My Harley FXD had its rear cylinder gobbled up by an especially ravenous one. I hope it gets 20W50 poisoning! :)


quicksand profile image

quicksand 6 years ago

Sorry it's "non capisco!" 20W50 poisoning is in chapter 2034, right? I am still on chapter 0003!

Anyway, 20W50 poisoning or not just take some precautions.

:)


Hal Licino profile image

Hal Licino 6 years ago from Toronto Author

20W50 is the weight of motor oil I used in my Harley so when the Xonu ate it, it slurped up all the oil. I hope the damn thing has been poisoned by it, or at least has a terrible tummyache... even though the tummies of the Xonu are in their 7th leg! :)


quicksand profile image

quicksand 6 years ago

You should take up writing science fiction as a hobby! :)


    Sign in or sign up and post using a HubPages Network account.

    0 of 8192 characters used
    Post Comment

    No HTML is allowed in comments, but URLs will be hyperlinked. Comments are not for promoting your articles or other sites.


    Click to Rate This Article
    working