Kantar Talks | Using artificial intelligence to measure earned media impact

KANTAR TALKS: USING ARTIFICIAL INTELLIGENCE TO MEASURE EARNED MEDIA IMPACT

Following this year’s Kantar Talks, we take a look at some of the inspiring and insightful sessions, with a specific focus on insights and measurement, and give our view on what they could mean for your business.

It wouldn’t be a marketing conference without the mention of artificial intelligence (AI) and machine learning (ML).

Historically earned media has been scared of measurement, but with the increasing amount of data and technical ability it is beginning to overcome this.

The industry is moving towards a place where it can uncover the true impact and exposure of earned media. In this case, earned goes beyond social conversation and attempts to quantify anything such as social, online content, print, non-advertised TV, etc.

However, earned media as a channel has various challenges which are not often present in traditional marketing as highlighted by Kantar’s Margo Swadley and Simon Ingram which include:

  • Actual reach is harder to uncover and often metrics used are potential or estimated reach
  • Sentiment of content can be both positive and negative making it hard to compare to traditional media
  • Prominence can vary vastly, and the context of the prominence can have a large impact
  • Planned Vs Unplanned, often earned media can be a product of virality or public occurrences so not always expected
  • Volume of content produced and speed at which it can spread can be far greater than traditional channels

Historically there has been a confusion between measurement and evaluation whereby simply measuring the estimated exposure of a campaign has classed as measurement, not the impact it has on a business.

By using AI and ML to truly understand reach, sentiment and prominence we are beginning to be able to get a better idea of the earned media picture across the majority of channels. By using these two broad techniques we can collect, analyse and process content pieces far faster than is possible manually. Not only this, but by training the approaches, we can work towards getting increased accuracy around both sentiment and prominence. This can then begin to be analysed more effectively and correlated with harder business metrics such as share price and short-term sales, putting earned media a lot closer to traditional channels when we look at it from a measurement and evaluation lens.

FUSE VIEW

Like earned media, partnerships across both sport and entertainment do not for the most part operate in a traditional media world whereby we buy ads and show the consumer our own creative. In-fact they can operate much more closely to earned media where the majority of exposures are generated through a property such as live TV, coverage across sport and entertainment publications and of course the vast amounts of user generated content across social channels in various mediums (AV, static images, text etc.).

Being able to accurately track the reach of our brand’s exposure across these channels alongside a better understanding of the sentiment and prominence will give brands better capability to integrate their partnerships into their measurement and ROI frameworks, and in-turn help them to justify their investment.

To read our full report on Kantar Talks, click here.