Part 3: The future of editorial analytics

So, in part 1 and 2 we've already discussed what Smartify is and how it works, now it's time to dive into where the future is headed and why Smartify plays such a vital role.

In 25 years, editorial analytics have come a long way. The days of naive publishing are over. Data in media environments has become more sophisticated, more real time and omnipresent and no established or aspiring brand knows how to live without graphs and numbers any more.

But if you take a stroll through these newsrooms, you notice that this addiction has brought its own set of problems. Editors are complaining that the data has become overwhelming and often hard to interpret. The technology and interpretation of data requires new skill sets that are not available in abundance. We have improved a lot, but after two decades we are beginning to learn that having lots of data does not directly lead to more efficiency or impact. With all our reports and dashboards, how are we going to take the next step? What is the future of editorial analytics?

Strategy and insights, data and tactics

Most modern media companies have known for some years that their data strategy should be a result of their business model and content strategy, and not the other way around. Choosing the right success metrics and measuring the connected audience behaviour patterns is only possible when the goals and plans are clear.

The art of translating high-level strategy into daily tactics - and getting from data to insights - is more widespread than even 5 years ago. Which kinds of stories should go to Facebook? Which are suitable for follow-ups? Where do conversions come from? Nowadays most editors are comfortable answering these questions with data analysis tools and they have the budgets to afford them.

Still, one crucial element is often overlooked. In the end, the results of a data inspired workflow do not depend on generic newsroom wisdom, but on the quality of decision making. Every minute, tiny micro decisions on an article level ask for editors' time and devotion. It is not enough to know which kind of stories do well on social media. It is crucial to know which specific article now, this minute, should be posted there. Just as data needs work to translate into insights, insights do not necessarily lead to the right action.

Actionability and decision making

It has become fashionable to claim that the data should be actionable, which means presented in such a way that it leads to result driven reasoning, and this is of course very true. But it is also a simplification, a 2D version of the problem at hand. A timely quality decision is not guaranteed by focused data visualisation: modern psychology shows that human reasoning is much too layered for that.

Firstly, editors are humans in a complex and hectic environment. It should not be taboo that decision making in such a world often becomes biased and dependent on the skill set of the editor at hand.

There is no energy to think about the relationship between the daily 1000 micro-decisions and the overall strategy, there is not even enough time to execute all the desired optimisations. In such a stressed workflow the model pyramid of strategy, tactics and action can quickly become an illusion.

Ask yourself the question: if you missed the chance to take action on something in your newsroom, was it because of the lack of strategy? The absence of a tactical plan? Not having enough time to analyse and execute? We estimate that around 60% of these missed opportunities are a result of the latter and few organisations seem to have a plan to solve that.

More fundamentally, we are used to the fact that analytics are a slice of history. But that is not how people normally make decisions. On our way out the door, we grab an umbrella not because our deep dive in past data has taught us anything, but because we expect that it is going to rain.

We take action because we see the near future, make reasonable inferrals about what is going to happen and decide how we want to respond to that. Similarly, if we expect editors to be result-driven and forward looking, aside from reports and real time dashboards, we should give them scenarios and accompanied predictions. Analytics should be more about the future and less about the past or present.

AI and the calculation of possible micro-decisions

If we follow this train of thought, the necessary innovation in editorial analytics becomes apparent: the task of quickly translating complex data sets into possible micro-decisions should be done by AI and machine learning. If they are ranked by predicted impact, humans can cooperate with the robot to make the best decision in a very efficient process.

Today's AI is not only capable of recognising the patterns, but also to predict the effects of the editorial action and to order the possible actions accordingly. If the company’s strategy is accurately translated in flexible and transparent algorithms, and the content strategists have total control over the settings, the AI can even assist in executing the most uncontroversial and simple tasks. But humans should always be in the loop.

The culture shift that newsrooms and content producers need to make is profound. It is not about the technology or data, it is all about reframing the basics. If we define the strategy of a media company to be the primary set of business rules and goals, tactics should be the translation of strategy into algorithms that apply to specific use cases to calculate scenarios and actions.

We have been doing that for years and we can tell you that the gains are great but this is not that easy. In coding and setting the algorithms you cannot escape everyday problems that were there to begin with, but until now we had the luxury - or audacity - to ignore.

For example, there is the ripple effect, that every action changes the future. If we A/B test a headline of an article on the homepage, it will influence the click-through rate of the headline next to it. If you factor in that impact, evaluation becomes less straightforward.

There is the fact that there will always be spatial scarcity issues, you can post only so much to Facebook, and articles cannot all benefit from distribution to primary spots. There need to be priorities, combined microdecisions. There will always be featured authors and topics and it will add bias to the impact data.

So, working with algorithms will align business goals and even teams and people, but it is not as easy as putting down some variables and reaping the benefits for years to come. It still needs to be a smart, dynamic process.

And more philosophically, the big advantage of historical numbers is that they don’t change any more. They have become a petrified fact. Predictions are possible facts and subject to change. It is even much worse than the umbrella and the weather. Analytics suffers from the Observer effect. You do not change the probability of rain by grabbing an umbrella, but you cannot act on the analytics without changing the numbers and especially the predictions. This requires an advanced mindset of all people involved, a knack for A/B Testing, experimentation and tweaking and systematic approaches to creativity. This is the true calling of data science in media intelligence.

There is even an important added effect that the smartocto team notices that every day; when a newsroom is confronted with smart and personalised notifications, tantalising predictions based on the algorithms they themselves have set, it is hard to ignore the content strategy. It will force you to rethink the overall goals and plans, to optimise and adjust them when conditions in the digital landscape have changed. The feedback even points to possible repairs in the company culture, and exposes the weakest links in the journalistic workflow that strives for maximum impact; whatever the success metric may be. That is the power of the cooperation of AI and people. And the newsroom will be the better for it.

In the age of artificial intelligence, we begin to see this pattern more often. AI is used to present the poker player with the best bets, even accurately predicting where a bluff can be most useful. AI can suggest negotiation moves and advise the control tower on the best ways to land a plane.

Especially when the data is complex and the decisions should be taken in an instant, AI will be the most trusted consigliere. Not to decide, but to make you aware of the options.

Smartocto’s Smartify and the role of data visualisation

The AI as an advisor, this is exactly what smartocto's Smartify is all about. The notification manager lets you translate the strategy into notifications and the system automatically accompanies that by machine learning predictions of impact to give you the priorities. The stream provides every user a filtered todo list, a continuous flow of possible micro-decisions based on proven tactics. The taskbox even shows a list of possible actions, filtered by timing and the editor's workflow. Ask the ‘what should I do now’ question, the system will match the data with the workflow instantly. That way you are 100% sure that the data strategy fits the biorhythm of the audience and the newsroom.

And to top it off, smartocto will close the loop. At the end of the week Smartify will send out a mail with the most important feedback learnings: how many and which notifications were done, whose experiments were AB-tested and what was the ROI of the total endeavour? Did the newsroom become more effective? With Smartify the cooperation of man and system becomes self-learning. And that is a paradigm shift in editorial analytics.

If notifications are the new main interface, the role of insights and data visualisation changes. Mail reports, dashboards with a lot of data points and filters luckily allow human deep dives in the data and creative pattern recognition by experienced data readers. We shouldn’t disregard that totally, the human mind is still a powerful tool.

But more and more, impact reporting - both historical and real time - becomes the validation of AI generated tips. It presents on what basis the system makes a certain prediction, it leads back to the ultimate ingredients of the recipe to comfort the users, to avoid the black box scenario, and to support strategic and creative thinking of the newsroom management.
The Insights no longer become the ‘What’ but represent the ‘Why.’ And this ultimately is the true meaning of the phrase ‘actionability of the data’.

Download the Smartify whitepaper