Dmitry Shishkin: There are lots and lots of editorial analytics tools out there and lots and lots of people out there claiming to know the Holy Grail of what makes content good, whatever you describe “good” as. It would be really interesting to understand, before we go into specifics, what problem is Overtone trying to solve? How are you different?
Christopher Brennan: Our tool is based on editorial signals. One of the key differentiators of Overtone is there are lots of different platforms who use — and there’s nothing wrong with this — use things like clicks and shares to decide with an article in terms of things like paywalls, ad placements, newsletters etc. Ours is based on the articles’ text, looking for what we call “journalistic signals”. So our algorithm can read through and say this type of story is doing one thing — it’s a short update about something in the news, it’s a kind of quick article based on a police report — and this article over here is doing something else, an in-depth feature.
Right now, you would need that article to be read repeatedly by an audience engagement editor or by someone putting together a newsletter to really use it properly. With our data, you’d be able to do that at speed and scale because you’d have that number sitting right next to it as metadata.
Reagan Nunnally: We are looking to give newsrooms back more time. We know they have so much going on, from research to editorial strategy to trying to talk to the business teams leveraging their content. And so if we can help in each one of those areas to give them just back some time, some breathing room, then we’re winning.
Dmitry: So all my signals are signalling things to me right now because I have been involved in understanding value for a long time now, and anything in that area really excites me. I’m interested to know, what can you do with that information? Obviously, you can potentially influence your publishing patterns. It seems like you’re going into the categorisation/taxonomy/metadata area in a really, really big way. In other words, is this another way of cutting data that some people already have or is it something like a new type of data you’re preparing?
Christopher: Sometimes, we use the phrase “another column in your spreadsheet” because we know that most newsrooms are data-informed. You have things like the tags on an article, and you have things like social engagements. But we are another column there that can help you make decisions based on those editorial signals and putting like articles together.
Dmitry: When discussing content analysis, I always refer to a triangle of different types of variables describing a piece of content, and one side of the triangle is a topic. The second one is a format, and the third one is a user need. If you only analyse topics, it’s pretty bland. If you analyse topics and format, then we are getting somewhere, at least from a production point of view. If you analyse that and user needs, then we’re talking about something more comprehensive. So how do you see that impacting journalists’ work?
Christopher: I think it’s more focused because it is about the editorial choice between these different articles. Again, we’re not saying that this article is valuable or another is not valuable: everything on our scale, ones, twos, threes, fours and fives, has value. It’s up to the editors themselves to choose what to put in front of their readers and when. That goes into things like personalisation. If a particular user likes in-depth articles about sports but quick hits on business, then that’s something that could be very easily done with our scores.
Dmitry: If, say, a newsroom operates five or six user needs, they know from the start, when they commissioned a piece of content, that it’ll have some elements. So what else does Overtone bring to the equation given that the ability to segment your content by user needs is already there? Because your categorisation from one to five is slightly different, isn’t it?
Philip Allin: It’s slightly different in the sense of what we’re really looking at, what we’re calling the depth scale at the moment, says something about the added human effort that was put into that article. So if you take two articles, they may be the same length, they may be written by the same person, published in the same paper, that doesn’t necessarily mean that they both have the same level of effort and work put in. And that’s what we’ve seen. That’s the sort of immediate distinction that we’re already able to make.
We’ve seen editors and editorial desks use that information differently. Whether it’s for a direct sort of monetisation, they want to figure out what goes behind the paywall, or what doesn’t go behind the paywall. How to improve your subscription and your readership engagement, for instance. Or I have an editorial need, rather than a business need, if I want to produce a newsletter or a widget on my homepage with some sort of top stories or something like that. There are different ways of using that same data.
But one way I’m looking at it is: this is really just the tip of the iceberg. What we do next and what we’re working on now is looking at not just the whole article, but let’s look at paragraphs in that article.
With that information, we can actually build up a fingerprint of the article and tell you more, give you more granularity, more detail about what that article is, and, therefore, how it might perform, whether that’s an editorial or business need.
Dmitry: We all know how wasteful newsrooms can be in terms of producing the wrong type of content. So, in terms of the actionability of your metrics, how do you see that it should be able to actually come to life?
Reagan: I see this as sitting at a keystone of a bridge. Editorial groups and business groups typically are very siloed. They look at totally different dashboards every day. We want this to be a common language between those two departments. And they can be a more cohesive group going out into the world and really feeling like it’s them versus the world, instead of these two siloed organisations. Our favourite conversations are with those that have had experience in the newsroom and start to realise that they are the keystone in the conversations that they can have with the business units. We want to put Overtone in their hands to provide that common ground.
Dmitry: So, talking about unique things, I would like to get your unique view as to where the market is going and what do you think will change in the next 10 years in digital publishing? What are going to be the big themes in the next 10 years?
Philip: What we’ve seen over the last 10-15 years, when digital publishing has become the de facto form of publishing, is a push universe. News gets pushed to your phone, to your screen, according to what a publisher, or an advertiser, wishes. And that’s been okay. That’s done a good job of democratising access to content and news. Everyone has gotten used to that. There’s a kind of an unbundling of the news products that you can see gaining popularity. And more broadly, I think, across the internet, users, people, and individuals are saying, “I want control over what I consume, and therefore I’m going to choose to pick what I want from where I can get it.” And I think news organisations and other content providers are slowly waking up to that fact and trying to figure out how to really reorganise their businesses to make sure that can happen and also can happen in a profitable way. Because obviously, we still want to be sustainable businesses.
Dmitry: What did Bill Gates say, “Most people overestimate what they can do in one year and underestimate what they can do in ten years?” What will not change in 10 years?
Reagan: For me, the reason I joined this project was really to bolster that human connection. We’re looking for heartbeats. We are looking for somebody with a passion and a desire to join the world of journalism and an equally excited reader to appreciate that. I don’t think that that is going to change over time. Even with technology as fast and furious as it’s coming, our technology supports that human connection in everything we do.
The post “Added human effort” – the most important column in your content analysis spreadsheet appeared first on WAN-IFRA.