Opinary Case Study: How to Ask More Engaging Questions

Opinary leverages a process as old as time (asking questions) with the theory that people have opinions more often than they have comments. We’ve learned a thing or two about what makes a good question. The definition of good? One that makes the audience (ourselves and you included) want to answer. The platform we use to ask those questions is Opinary’s primary product: a polling widget with an average engagement rate of 18 percent. That’s about one in five people leaving their opinion where one in 100 leave a comment. Here’s an example from PRI: Last year, two extraordinary Opinary interns, Rosemarie Foulger and Matthew Baughman, examined 1.2 million opinions shared on 923 Opinary polls. Using a variety of statistical methods, especially Genetic Matching, Rosemarie and Matthew reported causal links between the features of Opinary polls and rates of engagement. Here’s what they learned.

What type of opinions the audience
Continue reading "Opinary Case Study: How to Ask More Engaging Questions"

A New Newsroom Metric: Reading People’s Minds

This guest post was originally published on Medium. OK, sort of. Newsrooms already try to read people’s minds but don’t have objective metrics for it. From the start of the editorial process (what does “the audience” want/need?) to the analysis of the results (why don’t they scroll to the bottom?), it’s a guessing game as to what users are thinking. But there’s a new way to bridge that gap and measure audience opinion like never before. I’ll get to that below. First, let me tell you about monkeys.

Your Audience is Human

Believe it or not, I know a thing or two about monkeys.

I spent August of 2013 studying primatology in Kenya. One of the key debates in primatology is around anthropomorphization, or ascribing human attributes to the animal you’re studying. One side of the debate goes so far as to say that a primate in the wild
Continue reading "A New Newsroom Metric: Reading People’s Minds"

How the ‘Inverted Impact Pyramid’ Can Make a Hyperlocal Story Go International

This guest piece first appeared on Medium. Read more about MediaShift guest posts here.

Every story that matters to one person can matter to one million. That is because every story read or seen by one million people started as a story that mattered to a single person. That story picked up editorial perspective and frames to climb the inverted impact pyramid, gaining audience and reach, to become relevant to a more demographically and geographically diverse audience. The higher a story climbs, the greater potential for impact it has.

The Inverted Impact Pyramid


The inverted impact pyramid is a model that depicts the way in which a hyperlocal story intended for a small audience can become relevant to ever greater numbers of people, leading to a higher potential for impact. It also is a model that shows how any national or international story is grounded in a hyperlocal community.


global news international
Continue reading "How the ‘Inverted Impact Pyramid’ Can Make a Hyperlocal Story Go International"

The Internet Should Look More Like Reddit and Less Like Facebook

This guest piece first appeared on Medium. Read more about MediaShift guest posts here. There are three types of groups that filter the news: journalists, developers, and the crowd. The first two make decisions on your behalf, with their own and their organization’s economic survival being the ultimate motivation with few exceptions. The third has no motivation. In fact, the crowd’s curation of the news is a positive externality of an individual’s like, up-vote or share. Taken together, those individual approvals or disapprovals of material make for the most altruistic filter of content online. It’s time that we begin to think of the internet not as individual streams of websites chosen for viewing by the end-user, but as a rushing river of information that is fed by those streams. That rushing river needs to be managed not by newsrooms and Silicon Valley but by the people.

Three Filter Internet Content Model