winter / 2016

The magazine of branded content
Intelligent Content
Juliet Stott
Why it’s extremely effective
Juliet Stott
Feb 1, 2017

Many websites are so “bloated” with content that they prevent readers from finding the answers they need, says James Mathewson, the Global Program Director for Search and Structured Content Strategy at IBM. He believes this leads to lost customers and business opportunities.

Mathewson advocates the need for intelligent content in marketing which reduces, what he calls, the “needle in a haystack” information seeking process that most users go through. Here he tells Content Magazine how to use intelligent content to create excellent customer experiences, and why content should always give the users what they need.

What is structured, intelligent content?
Most content is unstructured and is in the form of text. It’s very slow for automated programs to analyze as natural language is very organic, changes all the time and has a lot of ambiguity in it. On the other hand, structured content has been created and tagged to match the audience, topic or journey state that the reader is in. Tagging adds context to the content and gives it meaning and purpose. Tagging content means you know who it’s for, its purpose, the journey state of the reader and what their next steps are. Programmed and machine learning systems can then easily surface more relevant content for that audience, i.e. recommend other pieces of content to consume. This gives the user a better experience, as it reduces the “needle in a haystack information process” that they are typically faced with.

How do you implement intelligent content?
There has to be a system behind the scenes that allows different content creators—authors, editors, production people—to know what their colleagues are doing. This helps them understand how they can work together in a governed way to build content for a common audience. In large organizations, like IBM, the biggest challenge is that the audience we’re trying to reach and engage with is finite. But, we produce content on the website as though the audience’s time and attention is infinite. The only way to break that cycle is to create a system where everybody has access to the data, which allows you to see what content is already available to avoid duplication. To do this you need to identify what content already exists, measure its performance in delivering information and then create links to it, rather than recreate it in a similar form elsewhere on the website.

What are the advantages of using data driven content?
The main benefit is to produce excellent customer experiences for your audience. Digital audiences are often overwhelmed with organizations producing the same functionally equivalent content. We’re cluttering already-bloated websites with information that’s hard to find. Digital marketers underestimate how time challenged and impatient the audience is. Cluttered experiences prevent readers from finding the answers they need. This is a lost opportunity, maybe a lost customer and gives an unfavorable brand impression for the audience. Using a structured content approach gives you a competitive advantage because readers can find the information they need easily.

Can you give an example how using intelligent content has enhanced a brand’s reputation, increased customer retention or driven sales?
We built the brand term cognitive computing from scratch back in 2011 using intelligent content. At that time, the term cognitive computing was Googled about five times a month, probably all by IBMers. As a company, we identified that this was a term we wanted to own. So, we set about compiling other related concepts for the brand—such as artificial intelligence, machine learning, natural language processing, algorithmic reasoning—and put them into a periodic table. We then built content around each one of the labels in the periodic table—answering basic questions of what these terms meant and how they worked together and including leading industry research on these topics. Every time we built content around one of the terms in the periodic table, we conditioned the conversation towards cognitive computing. Eventually this content started to influence the hearts and minds within the industry. After six months, there were 1,400 searches for cognitive computing in Google each month. Because we owned the word, the cognitive computing page captured the lion’s share of those clicks. It was full of links to all the other pages we’d created. There are now 2,000 queries a month on Google for cognitive computing.

How do you measure intelligent content?
There are many different ways to measure content. Using keyword intelligence, we’ve built a system to measure our content. This system acts as a feedback loop. It enables us to iterate, improve and find gaps in our overall content footprint. We also use search as a primary way to understand our audience. With the advent of voice search, more and more people are asking a question when they search. Even if they don’t ask a question, we can associate their queries to the questions implicit in the queries. We then mine those questions and rank them by volume and relevance. Next, we create content that answers those questions in the most compelling way we can, focusing on reducing the effort needed to understand the answer. After this, we measure the content – looking at how well the audience is interacting with it – what they’re reading, scrolling, clicking and which assets (a video, a demo, a trial etc.) they take away. Again, this measurement is a feedback loop, helping us iteratively improve the content. We can also identify repeat users from their digital fingerprints, so when they come back to our site, we show them new content, something that will help them take the next step.

Does data negate the need for a human editor?
Data helps you to understand the audience well enough to be able to present them with product-related content that they need to make a purchase. It’s about giving the audience what they need. Data isn’t necessarily important in storytelling, which is a current content marketing trend. It’s more about how to connect with the audience on a visceral level. We conduct sentiment analysis on social, so we can understand the big pain points for our audience, which we address with a compelling story or case study. But marketers have to realize that most of their audience are time challenged and they don’t have time to spend reading through stories that are not relevant to their information tasks.

In your book, Outside-In Marketing, you say marketers should use data to find their voice. How can they do this? What data should they be looking for? The voice you should find should always be relative to the voice of your audience. When people think about novels or fiction, the writer writes what he or she thinks is important, and the reader can go along with them or not. This is not the information model for digital content. The voice in digital content marketing must always be geared towards the audience. You need to differentiate yourself from your competitors in the way that you communicate with your audience, and that’s your voice. To do that you need to look at everything your competitors do digitally, and see how they’re creating content to meet their customers’ needs. This will help you not only understand how you compete for audience attention, but how you then differentiate your message to help your audience understand your unique value proposition. This is how you use competitive intelligence to find your distinctive voice.