Beyond pageviews. Attention, trust and metrics

by Andrea Iannuzzi – translated by Roberta Aiello

photo by Alessio Jacona from #if14 – The Whole Pic

Among the many insights that emerge every year at the International Journalism Festival in Perugia (here is the channel with the videos of the different panels), the theme of metrics to measure the value (not only economic) of online content deserves an in-depth analysis.

This issue was discussed in a panel I had the honor to participate in, with Pierluca Santoro (who has just ended his experience as temporary social media editor of the Italian daily La Stampa and runs Datamediahub), Gregory Galant, CEO of Sawhorse Media, Steffen Konrath, the inventor of Liquid Newsroom and Tony Haile, CEO of Chartbeat.

An excellent summary of the positions that emerged during the panel is contained in this article by Alastair Reid for the website, which saves me the effort of synthesizing the lively discussion.

I would like to analyze some of the concepts that have developed and consider in more detail a provocation that I launched during the debate, of which I am more and more convinced. Social media are not a tool for distributing one’s own content, it is wrong to use them in that way and therefore to give too much importance to return traffic to one’s own website.

1. Attention vs. trust

This is a hot topic when metrics are discussed. In Perugia, Tony Haile insisted on this point. If we want to give economic value to our online activities – selling advertising to advertisers – the only currency we have is the attention of the user, which is measured in terms of “active time” dedicated to content.

Others, such as Eric Scherer and Wolfgang Blau, argue that the real trading currency is trust.

As I said during the panel, the theme of the right metrics does not have a single answer because it depends on which “stakeholders” we are addressing.

For advertisers, marketing people and agents who sell advertising, pageviews still have an important value, at least because it is a metric that until now has been recognized and accepted by all. Although with pageviews we do not measure readers but just links and even though there are, more or less, unconscious tricks in order to increase the number of clicks, being able to count on multi-million monthly pageviews is a good business card in the market.

For journalists – in the broadest meaning as producers and distributors of content – the most valuable metrics are certainly more based on attention, ie on the time spent by the user to read, analyze and understand the content.

These are the kind of metrics that Haile refers to in terms of “currency” ie. of money, of revenues. Of course, in this case too, it is not enough to count the minutes during which the browser window is left open on the same content, because many multitasking users, in the meantime, engage in other activities – online and offline – that do not coincide at all with the attention given to an article.

Those who deal with metrics know that it is possible to measure attention through other parameters: the number of times the mouse is used, how much “scrolling” is done on the page and other markers of interactivity between the user and the displayed url.

There are people, like Jeff Jarvis, who reply to this theory affirming that the amount of time spent on content is not necessarily a sign of a good service offered to the users, who may instead need to find information quickly and can only be satisfied by arriving at their own goal in the shortest possible time (which is what happens to all of us with Google searches).

Haile agrees with Jarvis but replies that if the users are satisfied with the service, they will be back soon, giving away another fraction of their time and attention. Whether they are long or short time/attention units, their sum will always be the value that makes the difference from the point of view of economic return.

All in all, Jarvis, who speaks of the need to measure the “impact” of online content, therefore moving from a quantitative to a qualitative analysis (although the available tools have not yet been finalized completely), reintroduces the concept of trust.

This is the most important value for the third type of stakeholder ie. the reader, the user.  Personally, I consider it the absolute “value”, although I agree with Haile when he argues that it is the most difficult to monetize.

The debate between Haile and Jarvis was crystallized by Mathew Ingram in this Storify.

2. The need for combinatorial metrics

In line with what has been said so far, the problem is not eliminating the metric of pageviews, but teing able to integrate it. To use a pun, you have to “have faith in the attention”.

Combinatorial (and predictive) metrics are the result of real-time and simultaneous analysis of several parameters – related to interactions on the website but also on the social and to the users profiling – some of which are already used to optimize the context over content.

In the panel, I used the “dark room” example. If we are in a dark room, we have three choices: to move very slowly, trying not to crash; move quickly with the assurance that we are going to crash; or turn on a light.

The metrics, or rather the data underlying the metrics, are that light: flame, torch or lighthouse, depending on their precision, depth, freshness and our ability to interpret, assist and follow without being slaves.

Today journalists – ie content producers and/or suppliers of contexts – who refuse, for arrogance or bias, to get help from the data to assess the real-time interaction with their readers, have decided to move in the dark: they will go very slowly compared to the others, or are likely to go and crash into a wall.

3. Social shares and advertising videos: the false myths

One of the most fascinating and, at the same time, frustrating aspect of the digital age – which in purely historical terms is considered to be in its early days – is the fact that some “truths” are affirmed and removed in a matter of months or years.

It was, once again, Tony Haile in this article for Time who gave us a rude awakening regarding not only the equivalence between clicks and reading – ie the lie of pageviews – but also regarding a paradigm established between followers, whereby the effectiveness of content should be measured by its shares on social networks.

The search of Chartbeat debunks both assumptions, with an average time spent by 55% of users on a content item of only 15 seconds; and because only a small part of the content shared on social was actually read before.

The other myth that seems to have been debunked recently is that of video-ads, the so-called “roll” advertising inserted before and after, in the top and bottom, left and right, and in pop-up windows. The New York Times calls them “The Great Unwatched“, explaining how even in this case it is content with a very low impact on users.

4. Social media, the ambiguity of the “distribution channel”

This is the most risky aspect of my argument and I am open to objections, criticisms and opposite demonstrations.

When we analyze the metrics of our websites, we are all used to checking the percentage of traffic coming from social media, particularly from Facebook and Twitter. If we see an increase we are satisfied, thinking we have done a good job for our community.

I believe this is the wrong attitude, for practical and cultural reasons (which are also connected to each other).

The practical reason. Social media are paddocks, private property even if there is a public attendance. If the web is the square, Facebook is a cross between a bar and a fair. Social media offer us their free sharing platforms, their algorithms, their “graphs”, their users as potential readers. However, it is an apparent freedom because Mark Zuckerberg is not a philanthropist – and if he is, he uses the money he earns through Facebook for philanthropy.

Raise a hand whoever, when managing the Facebook page of a news website, in recent months has not seen the traffic – and the range of interactions – getting on the roller coaster, with peaks and rapid collapses. The fault is with the newsfeed algorithm, which selects – according to a criterion of presumed quality, penalizing the viral content – to whom and how many users show a certain content.

Complaining does not help, because Facebook is not meant to be a showcase available to a publisher or a company. It was created and lives to promote social interaction between “equals”, rewarding the most active and those who establish closer, stable, lasting relationships. Of course, there is always the option of sponsored posts. If you pay, there is an exponentially increased chance that your content will reach a greater number of users to whom it will be made clear, however, that they are sponsored posts.

Having a high percentage of one’s own traffic depending on social networks means holding an unknown constant on one’s own business. If that social changes its rules – or closes, or goes out of fashion – a large chunk of audience goes with it.

The cultural motivation. The function of social media is very different from that of traditional media, whether analog or digital. The word ‘social’ connotes to them much more than the word ‘media’. They are not created for broadcasting, but for relationships, for bi-directional sharing, to search for similarities, for being a community on the move, but not permeable to all those who seek to take advantage of it.

I have no scientific basis for affirming it, but I think that “using” social media as platforms for content distribution means confusing one place (for example Facebook) with a medium. In the long term the effectiveness of this form of broadcasting, as accompanied by the subsequent discussion with users, threatens to turn us into unwelcome guests, like a petulant door-to-door salesmen, who ring all the bells with insistence until a potential customer opens the door. With what spirit?

If we think about it, who else beyond the media tries to “sell” (not in the sense of economic value) products directly on social networks? The Facebook page of a singer is not the place in which the record label publishes links to iTunes. It is where the community of fans gather to share a passion, exchanging gossip, living together, knowing that there is something to share.

The recent “social campaigns” of some big companies do not relate to the products, but to the user experience. Why then do newspapers not understand that making “engagement” does not mean publishing news on a Facebook page and asking readers for their opinions?

It means, for example, asking our community to help us better understand a problem that we would like to deal with. It means trying to respond to questions and concerns about our alleged bad service. It means using our authority – if we have it – to ferret out the power where it is not giving the answers that it should give to citizens. It means making our fans or followers protagonists of the act of journalism as much as we are.

Therefore, is it forbidden to share content on social media? Not at all, but on two conditions. First, it must not be out of context. It is one thing to publish an article on salaries for teachers on its own Facebook page expecting that the “people” will read it as if it were a virtual newsstand. It is another thing to gain access to a community that revolves around the school offering, as a starting point for debate, the same article on teachers’ salaries. It is what can be defined as a kind of “Copernican revolution” from a “pull” mode, ie the newsstand mode in which it is the reader who comes to us to take the content; to a “push” mode, in which we try to offer targeted content to people.

The second condition is that we, producers and distributors of content, are people; the better the reputation of the people inside and behind that brand, the better that brand will function in the media marketplace.

If you’ve had the patience to get this far, I will summarize this long and boring post in a tweet; The right metrics are those that deal with people, not with users.