Week Six Readings: Web Analytics
Tracking traffic on the Web is tricky business with serious implications for those that rely on accurate traffic numbers for their revenue. Which of the methods presented in the readings on Blackboard seem most reliable and where do you see the effort to bring uniformity to the numbers heading? Note: The articles are posted in Blackboard/Course Docs/Week Seven/Readings. Please post comments by class time May 12.
May 12th, 2008 at 2:10 am
Great, I’ll post first. Blah.
To segway from Lessig, is there Read-Write advertising out there? Does that make any sense?
Moving on, I shall shill for Competitive Data.
The concept of “ecosystem context” is compelling, particularly because most modern journalists are more isolated than ever before. Also, as the readings showed, no one leading analytic tool exists, and without it, a combination of ISP-Based Measurement and Panel-Based Measurement seems to offer the best of both worlds.
Panel-based, though invasive, gives a good picture of what people are looking at it, and I don’t think the incentive changes the data, because the individuals aren’t surfing in a lab, but at their own leisure at their homes or offices, no?
Per ISP-Based, I am curious how close something like Google Analytics comes to Hitwise (also - is Hitwise cost prohibitive?) Solid search-data and psychographics from Hitwise would be a nice compliment to Panel-based.
May 12th, 2008 at 8:52 am
The best method out there now is using javascript plus what the author calls ‘web becons.’ Every major analytics package uses this method. Few use web logs anymore.
That said, I thought the Kaushik piece was confusing. There are a few important things that anyone should know about tracking visitors or customers.
Customers usually don’t like being tracked, at best they don’t mind. They can prevent you from tracking them very easily, and there will always be a chunk of your visitors who will not provide you with useful usage data.
The World Wide Web was not built with tracking or identification in mind. Most tracking systems are work-arounds or hacks. Javascript tags make collecting useful information possible, but its not perfect.
That said, the web does allow for some amazing tracking data. I think companies like Google and Amazon are going to rule this space. They have put a lot of money into tracking user habits automatically and compiling databases of interests. Nielsen and Hitwise collect web marketing data on the web, it seems, the same as for old forms of media: individual volunteers. I’m sure google has enough on me to know who I am, without me even knowing about it.
May 12th, 2008 at 9:03 am
I think with online audience research, as in traditional methods of audience research, people don’t really care if someone is paying attention to where their eyeballs and dollars are going, as long as it doesn’t annoy them, cost them money, or interfere with regular activities.
You sign up for discount cards at the grocery store because it’s unobtrusive and you save a couple bucks a week. But I hate the piles of circulars that I get in my mailbox, stuffing it full and making me miss my real mail a couple of times.
I mean, I don’t care if something is tracking which blogs I read or that I spend a ridiculous amount of my online time looking at lolcats. For the most part I tune out the personalized ads that follow me around, but if I get a pop-up ad, I definitely will not be visiting that site again.
Every once in a while I get a little paranoid about people tracking me, but for the most part I feel like it’s like a kid brother: Don’t bother me too much and I’ll let you follow me around for a while.
As both the readings mention, the field of Web analytics is still rapidly developing, but I think the best tool will be one that can gather comprehensive data without interfering in the surfing patterns of the users.
I think competitive data is the best way to analyze your audience, but something like Hitwise or Google Analytics, running in the background, collecting information about a broad audience will be more successful than panel-based measurement. This bothers me as it pinpoints specific users to track all of their web usage, and I think that these panel member’s habits would be different as they are aware that they are being tracked.
May 12th, 2008 at 9:46 am
After reading the piece, I can’t honestly say which method would be best to track user data. Which is why I plan to stick to journalism and leave the management of the Web site to the IT people.
On my blog this week I posted on a Wired analysis regarding the difficulty people have distinguishing what kind of data is traveling the Net (for instance HTTP, FTP, P2P). Better tools need to be developed, but I cannot honestly say what those tools should be. However, it seems the net is already ahead of TV in terms of audience data. Many still rely on personal journals in which participants write down what they watch in a given day.
I think the hesitancy advertisers exhibit when it comes to the Net reflects the much higher expectations they have when it comes to user data. Companies are comfortable advertising in newspapers with little more than circulation numbers to justify their ad dollars (how many people who buy or subscribe to a paper actually read it). Yet, with the Net they seem much more wary.
May 12th, 2008 at 9:49 am
I hate grocery store discount cards.
I also hate red light cameras, banner ads and web beacons.
So, I do my best to avoid them. I use the Adblock Plus (https://addons.mozilla.org/en-US/firefox/addon/1865) Firefox extension to prevent banner ads from loading, and I only allow specific web sites to set cookies. (By default browsers allow most anybody to set cookies.)
It’s not DoucleClick’s (http://en.wikipedia.org/wiki/DoubleClick) business where I go on the web. My actions on the Internet should be anonymous, unless I decide otherwise.
(Also, I don’t have a discount card, and I’ve never had to pay extra. Just ask. They’ll swipe one for you. But it looks like I’ll have to run for mayor to get rid of those cameras…)
May 12th, 2008 at 10:01 am
I agree with Adam.
Also, I’m not sure it matters. I understand, as “The State of the News Media” explains, that it’s important to have one standard way to track audience usage, whether it’s by unique visitors, time spent on a page, number of pages visited, etc. However, the Internet has proved itself to be capable of incredibly fast change and innovation. For every counter that is designed, there’ll be a site like Yahoo! Finance that uses Ajax and makes a certain measure useless.
It makes sense that advertisers want concrete numbers when it comes to the Internet, as opposed to print ads in newspapers in magazines. The data is possible to obtain, after all. Also, there are fewer newspapers: in 2000, there were about 1,500 newspapers in the country, and all came with quantifiable data: circulation. There are who-knows-how-many more websites. Advertisers wouldn’t want to take a chance advertising on most of them, unless they have proof that the website reaches a sizable chunk of their target audience.
May 12th, 2008 at 10:34 am
I’m confused! I mean, I can barely keep up with the tracking we do in class and stuff. Although, I will admit that I’m a little enthralled by my Google Analytics.
Like Adam, I can’t really pick out which would be the best method, some of it seems too complicated. I’m sticking to being a writer/journalist too … leave the Web management to my IT boys (and girls).
It kind of seemed like a combination of them would be the best. I mean, a lot of them had some really great benefits, I thought. It seemed like if it was good for one thing, it wasn’t good from an advertising standpoint, or not from a marketing standpoint.
The best tool would be one that is able to combine the benefits of all the currently tracking methods so that it’s benefitial on all levels, and not just one or the other.
May 12th, 2008 at 12:14 pm
Naturally, I am going to have to agree with/defer to the IT guy who posted earlier (Thanks for providing fodder, Ryan!).
But even though it is the most widely used method now, Kaushik shows some of the inconsistencies with the method:
“The data captured is less deep than twith other methodologies, but for tageted narrow purposes (banners, emails, and so forth), it works well.” Even though it isn’t incredibly precise, it still provides enough data to understand the types of content the reader is looking for.
In line with Bunmi said, I too am way too interested in seeing who is viewing my blog, how long they are viewing it for, and where they are viewing it from.
Despite the Internet being characterized as an anonymous medium, it is becoming more transparent with programs like these.
ps Canada loves Chicago sports.
May 12th, 2008 at 2:56 pm
As a former media buyer I am thoroughly confused by the lack of verifiable results for online traffic. While I understand that a longitudinal panel may be necessary to provide demographic profiles of each site I don’t understand why accurate traffic numbers are so hard to determine.
If Google Analytics can measure the traffic to my blog, why can’t Nielsen or comScore? Also, online advertisers and website providers argue over which is more important, unique visitors or pageviews but that issue depends on what an advertiser is trying to accomplish by using the website. Is it simply to reach the greatest number of people? If so, then use unique visitors. If repeated exposure to your ad is important then look at page views. The same logic can hold true from the content provider’s point of view.
It is good that the industry seems committed to increasing the accuracy of its traffic measurement methodologies. My guess is that neither source of traffic data, the website or Nielsen and comScore have the actual traffic numbers correct. But the rankings of websites relative to one another are probably usable by advertisers in terms of judging the value of a site and pricing it relative to other purchase options.
May 12th, 2008 at 3:02 pm
I like Bunmi’s points, especially as someone who was at first also a little bit obsessive about exploring google analytics (especially the map overlay. It’s so exciting, for some reason, to log in and see that you’ve picked up more readers in a new country).
And I also agree with Jen’s comparison with an annoying younger brother. I guess I, and I think a lot of other people who don’t understand the technology behind the internet very well, just kind of assume that everything I do can be (and probably is being) tracked by someone somewhere. And this doesn’t really bother me, mainly because I don’t have much of a realistic conception of how that information could be used (and who would really care that much what I’m doing).
So I appreciate being able to track things like the search terms and keywords that lead people to my site. It helps me gauge what kinds of content are effective and seems pretty unobtrusive for the readers. Because the point is to improve the product that readers are consuming, and they don’t even have to fill out a feedback form or anything.
May 12th, 2008 at 3:10 pm
Online traffic is such a revolving door at this point in the time that it’s either years away from having measuring gold standard or that the foundation of the web just was not built for such a thing. How many new favorite sites do we find in a year? How many sites that we spent most of our time on five years ago do we spend the same amount of time today? Political consulting has seen success in this sort of data mining with microtargeting. Another site to add to the golden geese of data collecting is Facebook. Although we don’t purchase anything besides a $1 novelty gift on it, Facebook’s personal and habitual online habits of its users are a gold mine to advertisers.
May 12th, 2008 at 5:53 pm
woah woah! Nielsen is cheating me out of 30% of my hits? That’s maybe one or two hits a day!
Ryan hits it on the head, the web wasn’t designed to track users, and for most people that’s encouraging. The old fashioned method used by the Nielsen people will not work long term since it cannot follow the power distribution method Clay Shirky talks about in “Here Comes Everybody.” (that a small percent of users will account for the majority of participation in special interest sites, skewing any “random sampling” or hit counting.
May 12th, 2008 at 5:53 pm
I don’t have much to add, but it seems like they must have figured all this out with television. How do they measure the difference between a family of eyes looking at a commercial or a single person sitting at home alone watching tv? It shouldn’t come as too much a surprise that the statistics all are different. There are lots of quirky ways to measure a site’s popularity. Time and experience will show us which ones mean more money.
May 12th, 2008 at 6:47 pm
I liked Adam’s comment about the fact that it’s strange how people are perfectly happy to advertise in newspapers based on nothing more than circulation numbers, and yet there is this much higher standard for web analytics. I understand that it has to do with the technology, but it’s definitely interesting that there is this need to know EVERY little detail and habit of every Internet user.
That being said, I find web analytics to be pretty fascinating. I mean there is simply no denying that it is interesting to look at the statistics and demographics of people who are visiting your site. As Christa said, it really IS kind of exciting to see when people from different countries have stopped by. It’s also interesting to find out the analytics of various sites that I regularly use. I certainly can’t blame companies for wanting to use analytics to maximize their service and profitability.
As a consumer, I kind of like when sites like Amazon remember my preferences and make suggestions accordingly. There is a certain element of this type of technology that can be a tad frightening, but generally, I like taking advantage of programs like that, and I think it’s pretty incredible that we even have these capabilities.