'Traffic index' claims a Google+ traffic drop. True?

By Sharon Machlis

"Failure to Launch: Google+ Growth Spurt Short Lived" claims a headline at Chitika, which published traffic data showing Google+ numbers declining after an initial surge of interest when the service moved out of private beta.

But what exactly is this data measuring?

Chitika's "traffic index" shows that Google+ soared the evening of Sept. 20, peaked on Sept. 23 and then started declining again -- for several days even dipping below the Sept. 17-19 numbers.

Unfortunately, there was nothing on the Chitika blog post explaining what the traffic index actually is.

"I should have posted my methodology," agreed Gabe Donnini, author of that blog entry. The index, he said, measures a site's "level of traffic activity versus the world" -- that is, how Google+ traffic compares all the other sources out there on the World Wide Web.

More specifically: "This number is calculated by aggregating the number of times a user was on a said domain prior to the domain at which they were shown an advertisement. This is a statistic which aggregates the number of times a user was on the site in question before they entered our network."

In other words, it's a more sophisticated way of measuring referral data: Someone needs to come to the Chitika advertising network from Google+ in order to be counted.

So here's my first question. Can we assume that referrals track overall Google+ usage? What if people are starting to use Google+ for discussions more than news- and link-sharing? Donnini countered via email:

Google+ is a social network in which the aims of the users on the site is to share information. If there were a community of users who were highly active within the site after they got there, wouldn't they be sharing and clicking on links, causing the count of Google+ referrers to rise? In this case we feel that a referrer distribution is an accurate proxy for the approximate level of traffic coming out of Google+

Perhaps. However, I think we need more information about Google+ usage patterns -- not to mention some official business profiles on the service -- before using this sort of referral as a sole measure of success.

I'm also not sure that Chitika's network is a representative random sample set for the overall Web, even though Donnini argued it is, saying it's got tens of millions of impressions and a hundred thousand publishers both large and small. What if Google+ users skew toward non-advertiser-supported sites unlikely to be in the network, whether academic, government or an e-commerce site such as eBay?

I'm fairly confident that traffic from Google+ to Web sites using Chitika's ad network spiked soon after Plus opened its gates and then declined from those peak levels. I'm not at all confident, though, that this single piece of data supports Chitika's conclusion that "Google+ does not seem to be able to drive unique visits in a sustainable fashion."

A snapshot of a week or two simply isn't a large enough sample set to glean long-term trends. Let me pick a dozen data points and the starting date, and I could show our own Web site's traffic is soaring or cratering. There are lots of short-term spikes and dips in many Web sites' usage.

"Given the huge amount of publicity and resources Google+ has received, why hasn't the social network taken off?" Donnini's blog post asks. Really? After less than a month open to the public?

Is my view colored by enjoying the battle for social networking users that Google+ inspired, since it seems to be spawning innovation all around? Perhaps. But I'm probably no more biased than an online advertising network trying to analyze the success of its largest competitor.

See more from the Data Avenger series.

Sharon Machlis is online managing editor at Computerworld. Her e-mail address is smachlis@computerworld.com. You can follow her on Twitter  @sharon000, on Facebook, on Google+ or by subscribing to her RSS feeds:
articles  | blogs .