How data and artificial intelligence are changing publishing

So what data should they be focusing on, and how should they use it to make judgements about the content they produce?

Steffen Konrath is the CEO of Liquid Newsroom, a company that uses technology to power its data-driven approach to content marketing, which it claims helps companies grow their B2B prospect and client list.

Here Steffen, who will be speaking at DIS 2017 on ‘why listening is so important to creating content strategies,’ offers insight into how data will shape the future of publishing. He explains why he thinks that the value of data is higher than the value of content, why technology will change the way publishers approach real time events and what role journalists will have in media offices powered by Artificial Intelligence.

About DIS:

Digital Innovators’ Summit 2017 takes place from 19-21 March (main Summit on 20 and 21 March) in Berlin, Germany. Now in its 10th year, DIS is a premium event attracting more than 600 top-level delegates from 30+ countries.

See more at innovators-summit.comRegister here if you haven’t. 

You have said that you feel that the value of data is higher than the value of content. Why do you think this, and what brought you to this conclusion?

Content serves a purpose and needs a specific context. It is usually only relevant at a specific point in time. Publishers want their content to reach the audience for which it was produced. They want to get it in front of anyone who is potentially interested in the information in the content so they can increase the number of people they can sell to advertisers.

There are however barriers which prevent publishers from getting that content in front of as many people as possible, and these can include; access limits based on location, the limitations of technology or just that the target reader may belong to a group of people not interested in content from a specific source. 

When I started to experiment with social graph analytics I figured out that such barriers can be made visible. So I decided to build a feature into my newsroom technology which assists me in exploring the hidden structure and characteristics of any information market. In short it is a feature which gives me access to the market DNA. 

So, for example, if publishers know who to talk to, what kind of content gets traction, what kind of people influence the decisions of others and how large the interest market is, they have the information they need to deliver any kind of content to any identifiable audience.  They can also target individual interests with the help of artificial intelligence, AI.

Once I figured out how to identify the DNA of any market, e.g. wearables, startups & VCs, and topical news markets, I knew I had something more powerful than the content itself. The knowledge about what kind of topics you have to choose to find a significant audience, and the ability to offer them content just by knowing them on an individual basis is, in my opinion, more valuable than the content itself. It helps you to identify the potential for new content products in a market by knowing your target audience by name in advance. It helps you to discover content gaps, for example when people in a market discuss topics of interest that are not being addressed by media or the industry. 

My newsroom uses content to serve an audience, to connect with people and to develop a language as a brand thereby creating a brand identity. However the knowledge about the markets is my core asset, it gives me value which grows over time and gives my company a competitive advantage.

Why do you believe that publishers aren’t using the full potential of data? How can they address this? 

What advertisers want is the best possible access to the audience of interest, and the only way to optimise the knowledge about an audience is by collecting and utilising data about them. So far high-end technology is used to optimise reach of sites (e.g. predictive technologies) and to sell inventory. 

But publishers can do much more with data. If you know the informational barriers in a market you can develop methods to optimise your reach. Data helps you to understand where in the consumer journey a current visitor is in realtime. You can use communication data to identify niches for new content products early in the process and monitor it to discover shifts of interest. This will enable you to make realtime adjustments to the products you already have in the market. If you build your website on personas and scenarios it is now possible to change the content of a site based on personalities of an individual person, instead of artificial characteristics of a group of people.

AI modern newsroom ()

How do you think that technology will change the way that publishers approach real time events?

Another exciting breakthrough we made was with realtime events. The Liquid Newsroom technology I use in my newsroom helps me to monitor events as they unfold. Listening to a market defined by interest in realtime opened a whole new world of possibilities for my team. I tested the process during the Content Marketing World in Cleveland (US) in collaboration with Communicate & Sell (owned by Ebner Verlag). The goal was to increase C&S’s  reach and awareness and to position them as a thought leader in the content marketing world. 

We monitored the communications around the event and then sent snapshot reports to the C&S team in Cleveland to assist them in identifying the topics which were driving engagement and accumulating interest. They then used this information to determine who they should interview. Thanks to the excellent C&S team the content production flow was perfect. 

Just imagine then the possibilities if you discover a hot topic and react to it with content in an instant. The campaign was highly successful and the joint approach helped to get the team and company to rank among the top three in mentions, engagement and impact at the end of the conference.

Using the same tech we can now gain a deep insight into the performance of communication efforts of sponsors and exhibitors at events. It is also possible to identify the prospects to which sponsorship packages could be sold in the future. If you know which topics are trending at an event you can focus on creating content that will connect with the attendees of the next event. So we use A/B test topics to help organisers plan for future events. 

What role do you think Artificial Intelligence is going to have in the future of publishing?

The overall volume of data that media companies access is set to explode. Think of IoT, the Internet of Things, and the kind of content distribution the availability of data from sensors will enable. In fact we are already exploring some promising ways to use artificial intelligence in combination with sensor technology in the Liquid Newsroom (LNR). 

I built the LNR not only to be able to post content to any of the distribution channels we know today (web, mobile, social media), but also to manage content on screens at physical places. That opens up the new field of contextual storytelling. 

Artificial Intelligence already helps us in our newsroom to understand if the nature of the content that is produced matches existing audiences or new users. The data we gather in the newsroom is the fuel for a neural network which helps us to discover patterns in user behaviour in realtime. The findings are not based on social media data alone. Patterns are used to optimise the content creation process, to serve our audiences with more relevant content and to manage distribution. With the ability to post to external screens, we can even use sensor data to find the optimal content mix. With today’s technology it is possible to recognise the age, gender and mood of a person. Imagine how powerful this could be if you use this data to adjust the content displayed to that person? AI can also harness existing data sources to enable contextual storytelling using such data like mood, weather, temperature or how many people have gathered at a location.

And also how will AI impact on retail? 

In a world which is increasingly connected (IoT), there are many platforms that can help users access information or buy products ranging from online to social to mobile and of course objects in the physical world. AI already powers search and recommendation systems today, e.g. Google, Netflix, and the retail industry can rely on such technology to recommend products based on data sources too. Sensors in retail outlets will contribute additional information to help increase the relevancy of the information provided. We only need to teach AI what to do with the data and to hand over the management of content to it, and it can automate most of the consumer journey. With AI’s ability to utilise the data of a single person, the segment-of-one approach becomes a reality. Super-personalised experiences will lead to higher customer engagement and ultimately sales. Its power lies in the ability to address microsegments.

What kinds of roles do you think that there will be for journalists in a future of AI-driven content creation? Will AI push more media brands into video creation?

AI will power part of the article creation processes to the point that human interaction won’t be necessary anymore. I discussed the impact of robo-writers on the work of journalists in a recent FIPP piece. With its ability to produce content at high speed it is much easier for AI to increase the sheer volume of output along the long tail of user demand. A common example is company reports, where AI could produce articles in the thousands, whereas the human powered newsroom would be slower and less cost effective. AI could also help to produce thousands of local content pieces, especially in areas, where data is the base – for example sport results. AI is highly effective when working within templates that enable it to create variants of content fast and easily. This works well for text, but for now we don’t have the technology to create video content targeting micro-audiences at scale and in realtime. But as we know from the entertainment industry, it is possible to use the technology to identify those factors which help a movie to become a blockbuster. While I think AI will lower the investment risk of producing expensive movies, it won’t push more media brands into video creation.

In the LNR newsroom AI is used to assist in the content creation process. With AI you can find meaning in large data sets, identify trends or the appropriate content to match the interest of microsegments. AI will also help to distribute the content to match the needs of a single person by heavy automation.

In such a world, the role of journalists will be to ask the questions an AI system will answer. It will still need that level of understanding a human domain expert has so they can teach a learning system what it should look for. And journalists will also still be needed to deliver insights based on their background knowledge and the ability to combine signals (e.g. facts) in a way which discovers hidden connections and reveals a new meaning (investigative journalism). Not to forget the human driven storytelling itself, which is also a huge driver of why we consume content.

Do you think that any publishers are using bots effectively at the moment? If not how could they improve this?

Most people think of bots as chatbots, but they are in fact new business platforms. Already some industries are moving their business processes to messaging platforms. 

If you use bot technology to allow your users to choose between alternative routes based on their interest, you face a new complexity problem. How do you collect data about its usage? How will you present it to advertisers? How do you monetise the advertising space in such a dynamic environment? We know how much companies, and not only publishers, are struggling in monetising mobile environments. With bots it won’t become any easier. 

If bot technology is powered by AI, we will have a greater ability to understand the individual user and to offer super-personalised content to them. But remember that bots can take over market processes (matching offer and demand) without the need of an intermediary. They have the ability to create – what I call – autonomous markets in which bots talking to bots (M2M) can fill your refrigerator with food once it is empty using data based on your consumption habits and information from the delivery side. They don’t however take into account a decision about what to buy that you made after watching an ad on TV. So how do you advertise in an environment driven by profiles machines have developed based on users’ behaviour or preferences?

Do you think that the age of Google search as a major tool for driving traffic to sites is almost at an end? If so why?

If your media organisation is running an eCommerce business you’ll sooner or later discover that relying on Google Search traffic alone won’t be sufficient to create the traffic for the conversions you need to reach your business goals. 

Google Search plays an important role, and will continue to do so until we see other technological platforms evolve and gain market share. But Google Search related methods, like SEM and SEO, are only one way to market content outside of your own site. The ability to advertise in front of an audience which has already articulated its need by asking a search engine to answer their question, makes Google Search a powerful tool for audience development. Also from a business perspective Google was the first company that enabled anyone to access demand curves for any product, service, or market of interest via its Google Trends service.

Facebook and Twitter offer new methods with completely different approaches. Facebook has gathered data on peoples’ likes, attitudes and opinions, which were previously only accessible via expensive primary research, qualitative and/or qualitative studies. However the company has not only done this on a massive scale, but also enabled the data to be collected and seen in realtime. 

Its social graph has allowed access to people behind people. In other words what one person is interested in might interest those connected to the person as well. This has means that marketing became much more effective because reaching one person often results in reaching out to many people via their connections. 

Another important innovation that Facebook has brought to marketing is that demand is no longer the single access touchpoint with a user. With likes, attitudes and opinions it is now possible to rely on probabilities if you want to advertise your products in front of a larger audience. And that audience is more likely to engage with your offerings thanks to predictive technology.

The launch of Twitter was another major step in disrupting audience development and marketing. In contrast to Google (demand) and Facebook (social graph), interest became the focus of the technology. If users became familiar with the service, they connect using hashtags (interest) to indicate the type of content they want to share. Previously companies had little or no data on how large their audience was. With Twitter it became possible to determine the size of your interest market, which is inevitably a much larger market than Google’s demand market. If you can identity the people who show significant interest in what you’re doing, you can try to win them earlier in their consumer journey.

What are the core services that Liquid Newsroom offers to publishers?

The technology and processes I use in the LNR are 2-3 years ahead of what is used in newsrooms today. Technology has inspired me to reconsider the processes of how content is produced and how facts are used. However a specific set of skills is necessary to make use of all the technology, and as most media organisations are struggling with their current transformation, I think even more changes would be too difficult to handle for them. This is the reason why my clients are usually operating in other industries.

My existing clients come to me to increase their reach, to grow their knowledge about their accessible market, to find content gaps, and to develop dynamic ways to respond to market changes in realtime. The data we retrieve from the Liquid Newsroom platform helps to evaluate market offerings to find the best context for their advertising. In the case of startups we can help them to develop a market before the product or service is ready, so that when they start to sell they are targeting an addressable audience, instead of having to start from scratch.

On the company side, I mainly work with the management teams on strategy. Much of the data I gather delivers insights on a highly strategic level. The unique approach I take contributes facts which assist in the medium or long-term decision process. Data allows my clients to understand hidden market structures or barriers. It can help to give influencer marketing and/or lobbying a new direction, and could even impact product development. If you listen to your market closely and in realtime what you discover might impact on service or product roadmaps as well.

Liquid Newsroom is a hybrid organisation. While it offers services or access to its proprietary newsroom technology on a license base, it is also a newsroom itself. The hybrid architecture helps me to finance R&D to test new ideas without having  to seek external investment. We can work pretty fast due to our agile processes. 

For me, DIS is a place to meet peers rather than new clients. I want to drive the discussion about the impact of AI in the publishing and the news creation process to help the industry to understand and make use of the next disruptive technology. 

Many of the themes Steffen discusses in this article will feature prominently at the 2017 Digital Innovators’ Summit in Berlin. For more information, click here.

More like this

IBM and Bon Appétit’s artificial intelligence cooking app

Artificial Intelligence unlocks the value of digital assets

Your first step to joining FIPP's global community of media leaders

Sign up to FIPP World x