1.8 billion reasons why image analysis will go mainstream in 2016

Consumers are sharing images online more than ever before. Already in 2014 people were posting 1.8 billion images on social media every day, which changes the function of the image for marketing and research.

Image-sharing platforms like Instagram have grown rapidly, now boasting 400+ million users and making it bigger than Twitter. Communication styles evolve to match the new medium: images are increasingly used as a means of expression, with text being consigned to a caption (often packed with emojis and ironic hashtags) serving the ancillary role of providing some, relatively necessary, context.

This means that, in 2016, the locus of meaning has changed: it’s very often more visual than verbal. What this means for consumer insight is that we can no longer rely on text analytics (whether quant or qual) to understand consumers through social media. In many cases we’re now dealing with the very opposite scenario: we have to analyze the image itself in order to understand what the (often sparse) text actually means.

2015 saw the commercial launch of algorithmic visual analysis tools on social media monitoring platforms such as our own Pulsar. Therefore, in this blog series I want to examine some of the emerging client uses cases we have seen that use large-scale image analysis. These use cases will cover three principles:

• Audience Understanding
• Trends
• Creative Development

To kick this off I’m going to discuss a use case on Audience Understanding.

Brief: Audience Understanding

A global spirits brand was looking to get closer to the digital lives of their urban, aspirational 24-35 year old drinkers so that they could make their new marketing campaign land more effectively with this audience.

They realized they knew a lot about who these people were overall – where they went out in London, what made them tick, and who they aspired to be. But when it came to applying that to digital campaign planning – ah. Where were they? How could they most effectively be reached? That was the gap we came in to fill.

What we did

We had two challenges to address:

1. Ensure we were analyzing the brand’s very specific target audience – it’s about pinpointing the core consumer, rather than anyone talking about them online. We wanted to reach a highly precise demographic and psychographic profile.
2. See the full 360 degrees of their digital behavior – including search, discovery and content consumption, alongside content creation and sharing. Traditional social media measurement would only give us the latter…

The solution? We created a full Digital Immersion research programme, combining the scale and real-world behavioural measurement of social, with the in-depth ‘whys’ of qual.

1. Bespoke, one-to-one qualitative style recruitment of a panel of the target audience to work with in this study
2. Tracked their social media interactions for a month across all platforms, gaining their permission to see their friends-only activity on Facebook and Instagram as well as Twitter
3. An online community, using mobile ethnography techniques to get our respondents to share with us how they searched for and consumed content – alongside the ‘whys’ of the things they chose to post and share
4. Social network analysis to position and visualise this audience in the context of their wider social media segments and communities
5. A workshop with the client team to download insights and build a playbook collaboratively alongside their digital strategy and media agencies

Why this worked

Image analysis of the most shared and engaged content from our social panel helped us profile the target audience’s perception of what their ideal aspirational lifestyle looks like. It also helped us understand the exact tone that needed to be used by the brand when communicating with them.

The content of the images showed us not only the topics they love (from street food and metropolitan living, to grime music) but more importantly their attitude to social acceptance when it came to the executional attributes (composition, filters, subjects) of visual content they would and would not share. Turned out, looking good in front of other people on Instagram was a big passion – and anxiety – for these urbanites… So we identified ways our brand could help deliver the both the kudos but also the credibility and reassurance they sought.

The study culminated in a working session where 10 guiding principles were agreed with brand and agency stakeholders, which now directs all digital activity for the brand.

*

In the next blog in this series I will be highlighting how we can use visual research to help identify trends.

Discover more about visual analysis in ‘Why mining images is key to understand social media today', or send an email to [email protected].