Companies typically like to hoard the data generated by their operations. It’s a natural defense: The less their competitors know, the better. But limiting the free flow of data can also limit the growth of a sector; there’s real value in knowing how markets are evolving, what’s working and what’s not, or just what the current state of affairs is.
That’s true for web analytics, the mountains of data generated by a website’s audience as it reads stories, follows links, and consumes content. Tools like Google Analytics and Omniture produce lots of interesting data, and nowhere near all of it needs protecting — there’s lots to be learned from seeing how one site’s audience compares to another’s.
So I’m hoping that, if you run a news site, you’ll contribute to greater information sharing by answering four basic questions about how your readers arrive on your site.
I’m prompted by seeing Josh Marshall of Talking Points Memo post yesterday about the growth in Facebook traffic to his site. In February, he said, 5.9 percent of visits to TPM came from Facebook, making it “the largest outside referrer to the site by a longshot.” Twitter, by contrast, generated only 1.7 percent. (Although I suspect he’s underestimating Twitter — see below.)
I’m also prompted by the discussion in the comments from Dennis Mortensen’s post in which he argues that news sites spend too much time optimizing for search (a.k.a. SEO) and not enough optimizing for generating clickthroughs from their front pages. His data, drawn from a number of news sites, found that nearly half of all article pageviews come from clicks on a site’s front page or section fronts.
Those are all really interesting data points, and I’m glad they’re being shared. But they’re by their nature individual data points. Just to speak from personal experience, here at the Lab, we get way more traffic from Twitter than from Facebook, and our front page barely generates 10 percent of our pageviews. Now, we’re a weird little news outlet — but everybody’s weird in their own way.
What I’d like to do is to increase the number of data points we have to look at. That’s where you come in.
I’m asking any reader who helps run a news site to answer four simple, straightforward questions about how your readers arrive at your site. They should be easily answerable with just about any web analytics software, but if you use Google Analytics, I’ve included simple instructions that will let you find the numbers in just a few clicks.
And let’s define “news site” broadly — whether you work for a small hyperlocal site or The Wall Street Journal, a small-market TV station or Gawker Media, a tiny niche site or ProPublica, kottke.org or CNN.com.
Here are the four questions we’re asking:
1. What percentage of your traffic comes from search engines? (If you use Google Analytics, click “Traffic Sources” in the left sidebar, then look at the percentage next to “Search Engines” below the line graph.
2. What percentage of your traffic comes from facebook.com? (In Google Analytics, click “Traffic Sources.” Look under “Top Traffic Sources,” on the left side. If you see facebook.com listed, it’s the number under the “% visits” column — you’re done. Don’t see it? Click “view full report” underneath it. Then, on the right side of your screen below the line graph, you’ll see “Views:” followed by six buttons. Click the second one, the one that looks like a pie chart. You’ll probably see facebook.com as one of your top referrers; if not, search for it in the “Filter Source:” box at the bottom of the screen. The percentage for facebook.com will be in the third column, under “Visits.”)
3. What percentage of your traffic comes from twitter.com? (The instructions here are the same as for facebook.com — just sub in twitter.com wherever I mentioned facebook.com.)
4. What percentage of your site’s visits begin on your front page? (In the left sidebar, click “Content,” then “Top Landing Pages.” On the right side of your screen below the line graph, you’ll see “Views:” followed by six buttons. Click the second one, the one that looks like a pie chart. Look for your front page, which will usually be called simply “/” — for most sites, it’ll be the first page listed. The percentage will be in the third column, under “Entrances.”
Post your numbers in the comments, along with info on who you are and what site you’re writing about. If you want to add some commentary at the end, great, but that’s strictly optional. The goal here is to get enough data to see whether there are any interesting patterns to pick out.
I’ll start off, to give you an idea of the format to use. Here are the numbers for at the Nieman Journalism Lab:
Nieman Journalism Lab
Period: Last 30 days
1. 15.76 percent of our visits came from search engines
2. 2.93 percent of our visits came from facebook.com
3. 12.93 percent of our visits came from twitter.com
4. 7.61 percent of our visits started on our front page
At the Lab, we’ve invested a lot in Twitter, where we have over 42,000 followers. (Much more than on Facebook, where we have about 3,000 fans.) And our audience is unusually interested in new ways of consuming information, so they’d disproportionately tend toward social media even if we didn’t treat it as an important channel. So our numbers are a little unusual for a news site. That means people find out about our stories largely through social media and links, not by coming to niemanlab.org and clicking around.
[Note: Just to head off a concern, I'm aware that counting traffic from twitter.com really undercounts traffic from Twitter. People using Twitter apps like TweetDeck and Twitterrific don't generate twitter.com referrers, if they generate referrers at all. But counting those app-using tweeters is difficult, so for our purposes, let's just use twitter.com traffic with the acknowledgement that it undercounts Twitter traffic. In our case, 12.93 percent of our visits come from twitter.com, but around 30-35 percent of our total visits come from Twitter in some form — apps or no apps.]
Just leave your numbers in the comments, in something like the form above. Thanks kindly, and here’s to hoping we see some interesting data.