Takeaways from Big Data & AI for Media, New York

First came data. Then it became ‘big data’. Now the data is becoming enormous. So enourmous in fact that most large companies have whole teams devoted to data in one way or another. I mostly work in gigabytes of data. Just a few years ago I was proud to have an external hardrive of 1tb holding my music and photos. Oh how naive I was! With IoT devices in so many places, the number of data points is incredibly immensely, now it’s about zettabytes.

And with all this data we can’t possibly process it ourselves so machines help. They are quicker and sometimes (often?) smarter than us. ‘Artificial Intelligence (AI)’ is a wide, often misconstrued, and sometimes contentious, term. Yet here I am, having used the phrase more in the past few days than I have in the last two years. Why? I had the pleasure of emcee’ing the ‘Big Data and AI for Media’ conference in New York and got to ask questions to the rather excellent line up of speakers that they had put together.

Questionning the Director of News at Bing during a panel debate

Focusing specifically on AIs application within the context of media makes it a little more manageable. Here are my key takeaways from the last two days of conversation focused around this subject:

1. Data is Empowering: the word ‘empower’ came up time and time again. From IBM Watson being the technology that empowers their customers to do more with data, to data teams empowering their colleagues in editorial & commercial matrixed organisations. Data empowers people to make decisions.
2. Hypothesize, test, analyze: this isn’t new but it seems that data is further supporting the lean startup approach for development.
3. KPIs v Context: most of the data sets being interrogated are owned, and we are using specific KPIs. But what if we’re not asking the right questions? As we trust the tools more, and as AI becomes more sophisticated it will be able to tell us what we don’t know.
4. Personalisation: organizations know more and more about their users, and their behaviors. Because of this, personalisation is feasible (and also a touch scary). Baby steps are being taken to test and learn how personalisation can be maximized for both editorial and commercial purposes without alienating readers and crating echochambers of information.
5. Sentiment: it’s not just about what we do, it’s how we feel when we do it and if it has the ability to change our mood. Sentiment analysis gives us a snapshot of how people feel about something which can be valuable, particularly for brands.
6. Public Perception of Data: consumers are starting to understand the value of their own data. This may change the playing field...
7. Privacy <> Value: it’s a trade off. As there is greater perception of data amongst consumers, they are becoming more aware of the value and what they are prepared to give. What if there was a world where individuals owned — and sold — their own data? And could this widen the rich poor divide as higher wealth individuals earn more from their data due to their higher propensity to spend.
8. Regulation: with the approach of GDPR in Europe and the Cambridge Analytica/Facebook scandal in the US, global regulation standards are seen as a ‘when’ not ‘if’.

From everything I’ve heard over the last few days there seems to be no downside to having a deep understanding of owned data. For media companies, data aides decisions around what to cover, to what depth, on which platforms and what time of day or week; it boosts subscriber numbers by understanding behavior and key triggers for conversion; and it helps advertisers target to the degree of understanding when readers and viewers have an intent to buy.

And whilst some media companies are ramping up whole teams and even spinning off data service companies, much of this is also possible with some of the excellent tools out there already: Parse.ly, Chartbeat, Heystaks, Amptitude, Unruly, NewsIQ, DNA, ARC, IBM Watson are just a few of the names I caught.

On a personal level it’s opened my eyes to the sheer volume of data points available on any individual. Whilst the conscientious media I’ve been talking to use anonymized data, not everyone will. I celebrate the potential data gives us — particularly around health — but am fully aware that it must be protected. European companies roll their eyes when GDPR is mentioned, it’s not easy to implement but data regulation is a necessary step in this journey.