While new digital outlets such as Medium and Quartz introduced the ability for users to 'annotate' or add 'notes' to specific content within an article, the New York Times recently unveiled a redesign which also brings comments beside the article in question.
Meanwhile, a group of "media engineers" have spent the past few years looking at ways to make engagement with an article a simple, yet effective opportunity for interaction between publisher and reader.
The result, currently being tested by around 40 media outlets – including ProPublica, The Stream at Al Jazeera and Fast Company – is ReadrBoard.
And while recent developments in the industry has given ReadrBoard an opportunity to "take advantage of the shift in expectations on the web around comments and interaction", the platform has been in development for a number of years, co-founder Porter Bayne told Journalism.co.uk.
"We have identified from day one that this is a design problem, so we're going to take all the time we need to take to iterate on user experience until we get something that seems to be resonating with multiple communities.
"And once we seem to hit that bar then we can start to think about trying to grow, trying to get on as many sites as we can and really building up the network that we want to build out to become an essential and ubiquitous feedback tool on the web."
And once applied to an article, the tool lets a reader select any piece of content, be it a word, sentence, quote, paragraph or image, and either choose a default reaction set by the publisher, or share their own.
If they choose to, they can also share a more in-depth comment, although co-founder Porter Bayne told Journalism.co.uk that this is "pretty rare", with "98 to 99 per cent" of readers choosing to just share a short reaction.
Leaving and sharing a reaction
When a user selects the relevant piece of content they wish to react to, and choose to leave a response, they will be presented with a series of reaction options, based on publisher suggestions, or space to add their own. These "custom reactions" make up around 35 per cent of responses, according to the latest statistics Bayne had seen.
Reactions, which tend to be either a single word or very short phrase, can also be anonymous, based on settings chosen by the publisher.
Bayne described the reactions as "the first, most atomic element that we think should be in community engagement", with the ability to comment further provided as an additional, but optional feature.
Publishers can moderate reactions and comments to varying degrees. Options include the ability to "autoblock all reactions", something Bayne said they "strenuously discourage", or insist those leaving their own responses sign-in to do so, also not highly recommended by ReadrBoard.
But he does "encourage" use of other features, such as the option to state certain words which will automatically have their vowels replaced with asterisks, as well as a function which lets you "block" specific responses from re-occuring.
To give publishers the ability to control the volume of reactions there is also a feature to "limit the number of bubbles that will initially appear", in a bid to prevent an article becoming overun with reaction 'bubbles' which appear at the end of relevant paragraphs.
"And we'll probably also roll out pretty quick the ability for a user to flip on and off whether these appear until you mouse over it."
Once a reaction has been left, this can also then be shared to social media.
If shared via Twitter, for example, the tweet will include the selected content, and a link back to the article. When the reader views that article via the link, they will be presented with a small pop-up box at the top of the article, which includes a piece of hyperlinked text to take them to the specific point in the article which the original reaction was based on.
At the top of the article there is a counter displaying how many reactions have been made in total, and when hovered over, a user can see a visual breakdown of the key reactions, giving an overall analysis of reader feeling toward the content.
The end of paragraphs where reactions have been applied also feature a subtle bubble logo, as referred to earlier, which again, when hovered over, reveals the reactions which have been applied within that paragraph.
The idea is that this makes it a much easier way for readers of the article to quickly get an idea of the overall response to the content, and the engagement which already exists.
"We can lay them out very visually, you open it up and even when there are 200 reactions on a paragraph, you can consume it as an individual in a matter of seconds and understand here's what the community thinks," Bayne said.
"I either want to engage further with that – mouse over and see what they're reacting to, maybe read some comments if there are any – or I can just move on."
Produce online polls
The tool also doubles up as an online polling platform, with publishers able to publish an article asking a question, again suggesting responses or allowing readers to add their own.
The results are then displayed in a visualisation, as shown in the example from ProPublica below, in the same way as reactions are collected at the top of an article.
"In terms of querying your audience to see what they think about something, it's another great way to get engagement," Bayne said.
Gathering reaction data
As well as simply looking at people's reactions to content, and using that information to help steer editorial direction, ReadrBoard provides publishers using the tool with data on how many reactions are being left on articles and how many views those reactions are getting.
According to Bayne, the statistics so far show that around 0.5 to 1 per cent of pageviews result in a reaction being left on an article, while around 10 to 15 per cent of pageviews are "reaction views".
And while publishers can use this engagement information to help with their digital strategy, Bayne said regardless, publishers using ReadrBoard should enjoy a boost in engagement.
"All of our publishers see an immediate increase in their metrics," he said, including the time readers spend on the website rising by around 10 to 14 per cent on average, as well as an increase in pageviews.
"And then correspondingly bounce and exit rates go down," he added, "so even if you're not looking at this information and making editorial decisions based on it, you as a publisher should see a positive improvement in your metrics."
Using 'reaction profiles' to enhance content recommendation
When it comes to revenue streams for ReadrBoard – and plans to introduce "premium services" in the future, such as "deeper" analytics – a wider idea is also to look at ways to use the insight achieved through ReadrBoard's monitoring of reaction engagement to offer an intelligent content recommendation service.
"We would rather use that value to increase their KPIs, do things like content recommendation, and have a content market, sort of like Outbrain, where people are paying to get content set in front of people based on what we think their reaction profile is," Bayne explained.