Web Metrics

We recently stopped using WebTrends for our Web analytics, partly to save some dough and partly to commit to other measurement services like Google Analytics and Quantcast.

I visited Quantcast the other day to see what kind of numbers our Web site was generating, having programmed our site to dump into their measuring protocols. I was met with proof positive that Web-audience measurement organizations still have a long way to go before they can be relied upon for accurate information, notwithstanding the frightening fact that advertisers use their stats to determine where to place large amounts of business.

Quantcast showed we had no measurable visits. Okaaaaay. In actuality, tcbmag.com has approximately 75,000 monthly visits and approximately 100,000 page views. Admittedly not Forbes.com numbers, but for a local business magazine Web site, not too shabby. Quantcast also identified what visitors we do have as mostly low-mid income females, a large share of them of Asian descent. Hmmmmm.

For the record, we e-mail a newsletter, twice a week, to e-mail addresses gleaned from our magazine-subscriber list and from those who’ve signed up on their own. I can tell you, having been one of the architects of the newsletter’s distribution, we weren’t concentrating much of our target audience on low-income, Asian females. Indeed, I’d hazard a guess there was nary a one low-income, female Asian sourced from our D&B lists.

Google Analytics, on the other hand, gets us about halfway to our actual tcbmag.com numbers, as measured by WebTrends over the past two years. GA says we’re getting about 50,000 visits at tcbmag.com and about 120,000 visits at sister-site mspmag.com. Close, but again—no cigar.

I’ve seen enough from magazine readership and audience-measurement organizations (e.g., Mediamark Research (MRI), Simmons, Readex, Mendelsohn, Erdos & Morgan, and Media Audit) to know they’re mostly accurate with time-tested methodologies.

Web-measurement companies like WebTrends, Google Analytics, Quantcast, and a whole host of others—including respected companies such as comScore and Neilsen—are by and large designed to be comprehensive and accurate. But right now, Web metrics is mostly a crap shoot—a parallel to much of the digital world.

What’s astounding is the amount of advertising being driven to Web sites based on highly questionable audience numbers. If I were a CMO at a major corporation, I’d dig out my framed statement from legendary retailer John Wanamaker about the uncertainty of advertising effectiveness: “Half the money I spend on advertising is wasted; the trouble is I don't know which half.”

Seems to me Web metrics is making it a lot easier to identify which half is wasted.

Comments

[Whatsername]

At least you're consistently cranking them out. My damn job keeps interrupting my flow of commentary.

[loose]

No, all those bangs and bings from you count towards your numbers. I on the other hand must rise and fall on the interest or neglect of my anonymous audience.

[Whatsername]

Wait - am I excised from the data?

[loose]

Thanks Andrew. Yes, we do exise our own internal "looks". Like the one I'm doing right now.

[AEklund]

Better late then never on this subject...

Discrepancies in web data more often than not occurs because of poor tagging of web content or poor set-up of your analytics account. Shit in, shit out, as they say. Also, and I'm pretty you do this over there at MSP, it's important to cancel out internal traffic from your reporting. You're typically voracious consumers of your own content so that needs to be taken out of the mix.

[BangZoom]

Holy Crap fellas...This is about metrics, not about media. This is about highly questionable audience measurement, not about whether the web is a viable forum for information and commerce. Jesus, read the GD text will you?

And incidentally, ain't no one showing me. I see it for myself and have since 1995.

[Jonathan Distad]

web advertising and marketing offers the most targeted and therefore highest ROI of any advertising available. Aside from qualified referrals, your time couldn't be better spent than developing your web and web strategy now because smart people are and once this economy starts to turn and the traditional budgets now allow for this it will either be too late or you will pay a fortune.

We use webtrends, GA and a number of others that provide a great picture of who visits client sites. Nothing's perfect but can you tell me who buys from strib ad or can I target best buy employees only with my ads in print media, no - online YES.

Those who say the web is still a fad or questionable form of marketing is obviously hiring the wrong ppl to show them.

[loose]

Tomasz - we had two years of reporting from WebTrends that actually matched up rather well with GA, thus indicating some level of integrity with both. Quantcast, on the other hand was off the charts incorrect, comparatively. Did you or Dennis see the seminal article in NYTimes Sunday magazine or the many other web measurement reviews posted on sites and in newspapers? The evidence is pretty overwhelming that web measurement is inconsistent at best. Although web advertising was up 19%, it's flattening, likely the result of a stinking economy. But also I wonder if the honeymoon may be moving into a next stage where "finding out" brings with it is a sense of disillusionment and second guessing?

[Dennis Rynearson]

I would welcome the other side of this discussion be allowed to present it's case. Web advertising is up 19% where print is down. I wonder why?

Also, I wonder was John Wanamaker's quote referring to print or Web. I think I know the answer.

[Tomasz Majewski]

While obviously the Quantcast information was anything but, may I humbly suggest that perhaps WebTrends was providing faulty numbers and not GA?