I recently came across a Wired.com article by @stevenlevy called ‘Can an Algorithm Write a Better News Story Than a Human Reporter?’ It’s an exceptionally interesting read for anyone interested in news, content, and digital media but, read it as a Content Strategist and it’s nothing short of mind-boggling.
In this post I give a quick overview of the article, pick out snippets that should get the hair of any Content Strategist standing on end, and hypothesize what this new technology could mean to the world of content strategy.
The article in overview
The article describes the work of a company called Narrative Science. Narrative Science (and their competitor Automated Insights) are leaders in the field of using computer power to automatically transform data into genuine and human-readable stories. Their work has been successful in areas of business and sports journalism where data is already rich and widely available, but they’re now also starting to produce content for the likes of fast food chains and best-restaurant websites and looking into producing personalised financial reports or synopsis on gaming sessions.
Although the stories produced are data-driven, they’re not necessarily dry and, as the article explains, functionality could be developed to include personal anecdotes from social media feeds or from human reporters. One has to wonder how long it will be before ‘writing engines’, as they’re being Dr Who’ishly called, become ubiquitous in a world looking for consistent, reliable, personalised, and cheap content.
The article expresses a sense that the ‘content bots’, or ‘writing engines’, to be more romantic, could know no bounds and, with quotes being bantered about like, “ultimately the sky is the limit*”, it seems that Content Strategists would do well to watch this space.
It all makes me think of a global architecture firm I worked with last year. Their mini-sized marketing team needs to produce and publish hundreds of project descriptions and updates which must be collated from thirteen offices around the world and presented in multiple languages. The descriptions are not meant to be emotive or spell-binding; they’re meant for architects and investors and should be factual, statistical, and to the point. If a ‘writing engine’ could pop out this content almost automatically, well, it’s an easy guess as to the rewards it could bring.
*Quote from Automated Insights founder, Robbie Allen.
Content Strategists, this should ring some bells
There were several sections in the article which were particularly interesting in terms of content strategy and, if you’re familiar with some of the core processes and deliverables, I think you’ll hear bells too.
Snippet no. 1
“…Narrative Science’s engineers program a set of rules that govern each subject, be it corporate earnings or a sporting event. But how to turn that analysis into prose? The company has hired a team of “meta-writers”, trained journalists who have built a set of templates. They work with the engineers to coach the computers to identify various “angles” from the data. Who won the game? Was it a come-from-behind victory or a blowout? Did one player have a fantastic day at the plate? The algorithm considers context and information from other databases as well: Did a losing streak end?”
Who are Content Strategists if not people who consider context and supporting information, then devise rules to govern subjects, and finally set down how the content should be formulated and angled for publication (usually presented in some kind of template)? Are Narrative Science’s meta-writers our more modern colleagues? The next evolution of content strategy?
Snippet no. 2
“…one of the meta-writers decided to build a story-writing machine that would produce articles about the best restaurants in a given city. Using a database of restaurant reviews, she was able to quickly teach the software how to identify the relevant components (high survey grades, good service, delicious food, a quote from a happy customer) and feed in some relevant phrases. In the space of a few hours she had a bot that could churn out an endless supply of chirpy little articles like ‘The Best Italian Restaurants in Atlanta’ or ‘Great Sushi in Milwaukee.'”
Sounds very much like the process and intent behind content templates. A content template is a document that presents (usually in table format) relevant components for individual content types, plus tips on how to produce those components. The primary purpose of content templates is to guide multiple content producers in producing content that is consistent and fulfils a particular communication purpose. Seems a pretty close match to the structure bots are being given.
Snippet no. 3
“Most news stories, particularly about subjects like sports or finance, hew to a pretty predictable formula, and so it’s a relatively simple matter for the meta-writers to create a framework for the articles. To construct sentences, the algorithms use vocabulary compiled by the meta-writers. (For baseball, the meta-writers seem to have relied heavily on famed early-20th-century sports columnist Ring Lardner. People are always whacking home runs, swiping bags, tallying runs, and stepping up to the dish.) The company calls its finished product ‘the narrative’.”
Narrative: now there’s a word straight from the Content Strategist’s most-favourite-lingo-list. Add to that the tasks of, “create a framework for the articles” and a choose a “vocabulary” and you’re practically a first class citizen of content strategy island.
Snippet no. 4
Now, to input a tone of voice:
“The Narrative Science team also lets clients customize the tone of the stories. “You can get anything, from something that sounds like a breathless financial reporter screaming from a trading floor to a dry sell-side researcher pedantically walking you through it,” says Jonathan Morris, COO of a financial analysis firm called Data Explorers, which set up a securities newswire using Narrative Science technology. (Morris ordered up the tone of a well-educated, straightforward financial newswire journalist.) Other clients favor bloggy snarkiness. “It’s no more difficult to write an irreverent story than it is to write a straightforward, AP-style story,” says Larry Adams, Narrative Science’s VP of product. “We could cover the stock market in the style of Mike Royko.”
When working with everyday people (not highly-skilled and paid copywriters) it can take months to make sure a tone of voice is being properly interpreted into content. Here you simply customise a bot.
Could bots give us more time for strategy?
A Content Strategist’s work relies not only on the integrity of their strategic decisions but also on how well those strategies are sold, explained, and implemented.
I spend a good deal of my working life trying to create processes and documents that will guide content producers in creating appropriate and consistent content. At the end of the day I know that no matter how good my strategy is, unless it’s implemented well my work will have little value. It’s perhaps this single point that leaves many a Strategist feeling that their work has somehow become less strategy and more project management. After all, if you’ve got a limited amount of time to get things done and the success of your work demands that your strategic suggestions are implemented well, you’d be wise to slice down strategic work in favour of a sparkling implementation. Rather look a winner on a few points than halfway there on a number. Speaking of winning. If a content strategy does fails to deliver the expected results, the first project element anyone with an ego will look to blame is implementation. After all, it’s far easier to handle the idea that content producers got it wrong than to find a fault in a much ideated and presented strategic plan.
For a moment, imagine the not-so-sci-fi possibility of Content Strategists as bot programmers becomes a daily reality. You may still need to sell strategic recommendations to stakeholders and continually educate your clients as to what it is you actually do, but you wouldn’t need to leave your strategies in the hands of content producers and hope that they’re able to glean from your workshops, content templates, and style and tone of voice guidelines exactly what your vision is. No. You would simply programme your bots and wait and watch as the website was populated as ordered, on time and on tone. You would have full control and, with bots taking ultra-consistent care of the content, the only opportunity left for your unique human genius to shine, would be in the strategy.
Botulism: the not-so-nice slice
This would be the cold and scientific world of perfectly parsed code where the only ‘bacteria’ in the process could be you. Leaving implementation to can-do-no-wrong bots would shine a very bright and clinical light on the worth of strategic decisions. If results were not according to plan and the content bot programming was right, well then, the fault could only be in the strategy. I personally think it would bring exciting times: we would feel more intensely the cold floor of our failures (and the warm light of our wins) and, with the tweak of a bot, we’d be able test and iterate out strategies at the speed of… well, at the speed at which we could sign-off our re-tuned strategy. Lean content strategy here we come.
To bot or to not
This all sounds a bit like a fantasy (or a nightmare, depending on where you stand) but auto-content is a real-life tool which, budget permitting, we could be making use of and suggesting to clients. So, Content Strategists programming strategies into bots? Puh-lease. Or yes please?