I chatted up a piece of new analysis from Nielsen around the office a good bit today.
In the analysis, David Martin, vice president of primary research for Nielsen Online, described an uphill battle that Twitter, the oh-so-buzzy status network, may face growing Internet market share while its user retention rates languish around 40 percent.
“By plotting the minimum retention rates for different Internet audience sizes, it is clear that a retention rate of 40 percent will limit a site’s growth to about a 10 percent reach figure. To be clear, a high retention rate doesn’t guarantee a massive audience, but it is a prerequisite. There simply aren’t enough new users to make up for defecting ones after a certain point.”
OK, interesting take, though Peter Kafka observed that Nielsen's counts may miss a lot of Twitter traffic.
“The majority of Twitter use happens away from the site, on mobile phones and apps like Tweetdeck, and it’s theoretically possible to be an avid Twitterer but never visit Twitter.com after you sign up.”
Regardless of Twitter's prospects, I found myself more interested in Nielsen's plot of retention rates, and the regression analysis with it. (Rather than scrape the chart to show it here, please follow the link in the first paragraph to see it, the first of two charts, in its original context.)
That regression line shows a distinct curved-line correlation between Web site retention rates and Internet reach. Yes, common sense would tell you that a site that brings 'em running back at least every month stands a better chance of growing share than a site with logfiles full of drive-by visits and one-hit wonders. And the inverse makes sense, too: the largest, most far-reaching sites online today must have high numbers of frequent visitors.
But the plot shows retention and reach do not correlate in a straight line. Not until a site achieves more than, say, 50 percent retention can its proprietors hope to grow Internet reach much past 10 percent.
So I place that analysis in the context of a typical local media site, which by no stretch should make it to the top right side of Nielsen's plot as long as overall Internet reach is the x axis. If, however, the x axis were Internet reach confined to that site's local market area, shouldn't a good media site be high on that line?
Yes! If only it were true!
But we know, of course, most such sites instead fall somewhere south of 20 percent local market reach, when reach of any companion media (printed newspapers or broadcast outlets) is taken out. We also know that managers of most such sites wrestle incessantly with a traffic power law curve, especially an outsized “long tail” of infrequent visitors generating short sessions and low page counts.
Poor retention, in other words — just like the characteristics Nielsen attributes to Twitter traffic. To me, the similarity makes sense: The consumer value of a social-status service like Twitter resembles the value of “news” as a service. It is incidentally important, but not always important, and never all important to any one person. The intervals between incidents that you or I might deem important defy any prediction.
Does that mean we in local media should hope Twitter finds a great business model to capitalize on its traffic, since it so resembles ours? Hmm. Maybe. I find it easier (though not completely easy) to connect those dots than the ones between us and, say, Google, Amazon, Microsoft, Apple or eBay.