Art And Science: Data Informed Content Creation

12/10/2016 12:16

Data and creativity; it's a relationship that's been debated for a while. Can the two truly work side by side, or is there a limit to how much we can push art and science together?

Measuring Creativity Works, But Has Its Limits

It's now commonplace for all major content outlets to measure the performance of their outputs. Being able to tag and test just about everything on a page: Font, content length, headline, video player size, measurement and optimisation has become much easier than in times gone by.

As such, we should firstly ensure brands are indeed learning as much as possible from the content we produce with them. Ensuring content is tagged and analysed should be a starting point at which we push clients forward from. The balance however lies in what you do with this data. Whilst a natural progression would be to continually refine to "spiking" content, this has two negative implications.

Firstly a state of continual refinement can result in narrowing your field of focus too much. Whilst it is granted that you can consequently engage with a core audience deeply, you can miss a lot of incremental engagement with consumers. As Nigel Gilbert of Appnexus notes* newspapers would lose their integrity and become gossip magazines if they only provided high traffic producing content. The responsibility is there to service the whole of their audience.

The second issue this over-analysis can cause is ironically a lack of creativity. Jim O'Donell from Hulu has said* that whilst analysis of previous shows helps them know whether to support subsequent series, relying solely on this means nothing is a surprise. The world needs genius. Unexpected heroes that sweep audiences off their feet. Whilst it's natural to develop content strategies, if you fail to test new things new consumer trends and behaviours can leave you behind.

The Programmatic Pickle

Whilst a huge topic in itself, data and creativity can't be discussed without including programmatic. With algorithmic TV buying finally a reality, automation is fast becoming the way of the world for everything. There are consequently natural implications on what this means for creativity.

Many are talking about programmatic as the answer to all of life's problems, in enabling brands to reach everyone with their content cost efficiently. Disregarding the fact that with now commonplace PMP whitelists, programmatic at a basic cost metric level now is often more expensive than standard buys, this is perhaps extending the abilities of the content activation method too far.

Though yes programmatic offers a way to reach audiences at scale, too often the environment in which we place ads is being forgotten about. As discussed by James Hayr at a recent IAB UK conference, context should be queen behind the content. Whilst reaching the right audience with the right content is a fantastic progression for the industry, if we're doing it at the wrong time and in the wrong place, you may as well not bother.

Networks and trading desks are understandably protective of providing granular insight into campaign performance at a site placement level. This is because often when examined, a pure audience buy focus produces poorer results than when context is focused on as well. As such though programmatic can indeed scale your content, a balance needs to be had in how you do so. It's better to reach less people and provide a meaningful connection with your content than to simply reach as much of your audience as possible and provide poor consumer experiences.


Creativity is by nature not 100% predictable, so whilst we can and should analyse and predict performance, data should be used more to steer us than definitively lead us in our thinking and actions. On the other side of the coin, whilst data and programmatic offer a great way for us to reach our audiences with the content we produce, we need to ensure we're not sacrificing quality and performance for reach and scale

So, can data work with creativity? In short, yes, but we need to be careful in how we use it.

By Owen Lee

*In DMEXCO 2016 seminar: The Technology of Storytelling Data-informed Content Creation: Unlocking the Power of Art and Science

N.B. Above is an outline of conclusions reached from attending the panel seminar The Technology of Storytelling Data-informed Content Creation: Unlocking the Power of Art and Science at DMEXCO 2016.