Why is it that most magazine apps fail to live up to their expectations?
One reason is that the majority of consumers’ time on mobile is being spent on just one or two big content distribution platforms such as Facebook and Twitter. With a rapidly expanding world of online content just one click away, are they simply trying to solve a problem that consumers don’t have?
The app I'm interested in has not come from a publisher but from ‘Big Blue’ who have used their artificial intelligence engine, Watson, to create a uniquely data-driven spin on a publishing staple - the buyer's guide.
In a recent report, Technology for Marketing showed that marketers are both interested in and ready to invest in content automation. It’s understandable. At a time when 1,212 WordPress blogs, 400 YouTube videos and 3.3 million Facebook posts are published every 60 seconds, finding ways to automate elements of content production is one way of giving brands and publishers a way to compete with the rising tide.
IBM Watson Trend is an intriguing example of what automated content might look like in an app.
The app uses natural language processing and machine learning to intelligently analyse millions of conversations across social networks, blogs, forums, comments, ratings and reviews. These information sources are curated into a real-time buyers guide, forecasting the products which will be ‘hot’ and forecasting consumer interest.
Launched at the end of last year, it fulfilled a real need. Namely, ‘What should I buy my son for Christmas, without having to, you know, talk to him?’.
The Register announced the launch in suitably snarky form with the headline, “It's come to this for IBM: Watson is now a gimmick app on the iPhone”, and article bemoaning that “billions of dollars in machine learning R&D concludes that kids like Lego”.
Well, I for one am pleased IBM have applied this technology to the consumer space. The product recommendations are independent and unbiased in a way that most magazine buyer's guides are not, based on data (and lots of it) rather than who has the most persuasive PR team.
The UX is very clean and easy to navigate, with trending products presented as tiled images with their trend index, a score out of 100.
The navigation clusters trends together into topics such as Tech, Toys and Health, and is generally a good experience.
On clicking into one of the trends, you are presented with ‘story behind the trend’ – a short overview of the trends. Scrolling down, you see ‘The numbers behind the trend’ – a graph which charts the Trend Score for that product over time, with a forecast of how it will continue to perform. The graphs are interactive, responding to being touched.
Underneath that is, ‘what people are saying’, a selection of quotes, snippets of conversations from Twitter, blogs and review sites. Beneath that is a selection of images of the product pulled from Flickr. It is a pleasing bundle of different content types which works well together as an editorial package.
My main gripe with the app when it launched was that it was not fast-moving enough to make it something to refer to regularly, as products such as ‘Samsung TVs’ (quite a broad category in itself) have continued to trend since launch.
This is something IBM are clearly aware of as they recently introduced 'Emerging Trends' at the front of the app to make it something to use more than once.
As far as predicting the future goes, it’s not particularly ambitious. With very broad upper and lower probability bands and forecasting no more than seven days ahead, it’s about as forward-looking as the weather forecast, and far less precise.
To become more useful, I would also like to be able to cut the data by demographic data – best products for motorsports fans, for example – and to search for particular products in a similar way to Google Trends.
But to me, these are just bumps in the road for what is a fascinating application of technology.
If we are to make apps that aim to push the magazine format forward rather than simply recreating static pages within an app, then a constantly updated buyer’s guide based on machine learning is a good example of what form that might take.
Another interesting application of this technology is IBM Watson’s recent partnership with Ted Talks. Watson has digested two thousand or so videos and is able to use that bank of knowledge to answer questions, such as “what is the secret of happiness?”
It’s intriguing stuff. Now, if only I could connect Watson Trend (so I know what to buy) to my Google Calendar (so I don't miss a birthday) and then to Amazon (so I don't have to go to the shops)… I might automate my way to becoming the best dad in the world.
IBM Watson Trend can be downloaded from the App Store.
