If you have been anywhere near the neighborhood of digital publishing in the last couple years you may have heard some discussion about quality content. In fact you might have heard quite a lot about it. It has been posed as the answer to any number of questions.
For journalists, it is their raison d’etre, what separates them from PR people, marketers, bloggers and the rest of us in general. The quality of their content is based upon context, journalistic standards, objectivity and so on and so forth.
Marketers on the other hand seem to have finally come to the realization that as they have focused on placement, marketing content as they have known it for decades is primarily garbage. Having taken the next step in analytics they have discovered that the video and banner ads they have sunk their money into have been largely ineffective. So quality content, in the form of content marketing or native advertising or brand publishing, is viewed as the antidote to the failure of display.
The quality quest is also a big topic of conversation at the big tech companies. Why do you think Google tweaks its algorithms several hundred times a year? Or why does Facebook change its news feed parameters so often that it hardly looks the same from one week to the next? Do you think writing code is the way to identify quality? I don’t.
What the search robot programmers use is a statistical analysis of attributes that correlate with quality. So keywords seemed a good predictor of relevance and links a sign of authority. These types of measures however were easily gamed, then overused, so Google at some point started associating quality with sites that didn’t use too many keyword repetitions or too many links.
The dataheads and their legion of followers believe that the answer to pretty much everything lies in analyzing ”big data.” Can big data find big quality? In a word, no.
A few weeks ago I met a guy who worked for one of the big three financial news services. He is a veteran journalist who works on the commodities desk. He expressed his frustration at a new management group which was encouraging him and his colleagues to include mentions of celebrities in their headlines.
If you don’t think too much about it, that is what the data will tell you to do. Mention Justin Bieber in a headline and you get way more views than a headline about the prices of soybean futures. Is that quality? How does that play with the readers of this guy’s reports who are likely commodities traders? Data measures popularity and that’s completely different than quality.
That example also shows quality is defined by the reader. One person’s quality is another’s trash.
I wrote a blog post for Beyond PR a while ago in which I attempted to define quality content. The short answer is that good quality content at minimum must be either interesting or informative. It must educate or entertain the audience, whether that is a mass audience or a small narrowly-defined group.
Deciding whether something is interesting or informative is not something robots, algorithms or data analysis do very well.