By: Keith Jordan
I lived in New York when Rudy Giuliani was mayor and one of his catch phrases stuck with me: “You get what you measure.” He didn’t originate the phrase, but he’s the first public figure I remember using it.
Giuliani was referring to crime and budgets, but the same principle applies to your publication. With digital publishing, you can have far more precise, specific ways of measuring what your readers do than you have available in print or broadcast. But there’s a risk of losing sight of what matters by focusing on the wrong metrics, or whatever metric is trendy at the moment.
When I first started out in online journalism in 1995, the only metric anybody had heard of was hits. A hit was one file downloaded one time. A basic Web page with no images downloaded by one user was one hit. This was a fairly valid metric when most sites actually were one Web page with no images. The problem was that a Web page with five graphics, visited one time by one user, was six hits (the page itself and each graphic). Early Web publishers realized this and started loading their pages with small graphics, rapidly inflating the number of hits.
How to solve this? Page views! You don’t count the graphics; you simply count the number of HTML pages downloaded. This takes us back from six hits, in our example, to one PV. But publishers found ways to game this stat, also; for example, by creating photo slideshows in which each photo was an HTML page.
You can point out flaws like this in every metric that has come along, even today’s gold standard, the unique visitor or UV, which makes no distinction between a heavy daily reader and someone who visited one time for seconds. The truth is that while each metric gives you a piece of the puzzle, no metric gives a holistic view of your traffic.
If I were running a content site, the metrics I would insist on seeing every day are unique visitors, unique visits, time spent per visit, visits from search, top social shares, top search terms, and percentage mobile/tablet. You (or someone on your team) can configure a dashboard in Omniture or Google Analytics to receive this information at the same time every day. If you aren’t used to seeing these stats, give it a month or so before you react to them. Otherwise, you might overreact to statistical quirks (for example, you get more UVs early in the month and not as many late in the month, due to the nature of how UVs are calculated).
A word on social: Tracking user engagement on Twitter is a whole separate topic. It is very difficult to know when a link to one of your pages has been retweeted. One of your articles could be getting lots of buzz on Twitter without your realizing it. But as long as users are sharing content mostly by clicking the share buttons on your site, you can track that by using standard code from companies such as AddThis or ShareThis. For other social sharing, including Facebook and LinkedIn, Omniture and Google Analytics do a good job of tracking the number of shares by finding linkbacks from those services that point to your site.
I hear lots of anecdotes from colleagues about the companies they work for either not tracking metrics on a regular basis, or not understanding the data. A former co-worker recently told me that when executives asked his division chief for some quick metrics about their product, they had to ask for a recess during the meeting because they didn’t know offhand. This is at a $20 billion media company. This same group had invested heavily in building products for platforms, including BlackBerry and Android, without any idea how many users they had on those platforms (very few, as it turned out).
If you don’t receive broad-spectrum metrics every day, put down this magazine and get a dashboard set up to deliver them to your email at 5 a.m. daily, and commit to reading them first thing every morning. Flying blind is no way to run a business.
Keith Jordan is managing director of Upstream Digital Media (upstreamdigitalmedia.com), a consulting company focused on digital publishing.