Across online publishing, the competition for advertising dollars is intense. It’s a tough battle. Falling cost per mille (CPM) is still eating away at the profitability of the industry. The migration of audiences to mobile devices doesn’t help. Mobile inventory has been notoriously difficult to monetise due to lack of data; the inability to track via cookies and fragmentation of the audience across apps and devices, which has resulted in low CPMs.
Today, more content is created than is actually needed. Anyone can be an online publisher and the ease of becoming one is driving down average revenue per visitor and further increasing competition. Increasingly, the only way for publishers to differentiate themselves is through attracting and keeping the right audience and more importantly understanding and monetising them effectively.
Granular insight
The key almost certainly lies in more sophisticated data and personalisation. The sheer proliferation of competing ad tech solutions doesn’t make it easy for publishers to distinguish between the sound and the unsustainable.
Yet gradually the focus is moving towards more personalised and granular insight derived from the capture and analysis of a range of data, including social which also reveals expressions of interest and opinions. Used wisely, this can uncover hidden affinities to reach a far deeper insight than was ever before possible. Indeed, a ground-breaking new study carried out by researchers at Stanford University and the University of Cambridge found that computers’ judgments of people’s personalities based on their digital footprints are more accurate and valid than judgments made by their close family, friends or acquaintances.
The truth is that publishers already have a wealth of first-party data at their fingertips. Most are in no doubt already tracking the most popular pages on their sites and using behavioural tools to understand the semantics to see which topics are the most popular. But a deeper insight is possible when this is used together with social data. If an individual can be pinpointed through a social login – this information becomes even more valuable as, if permission is provided, they can be targeted directly with information that reflects their interests.
When a publisher has access to a reader’s likes and interests, certain natural clusters emerge which are quite clearly defined – in some cases down to a certain football club or a certain fashion brand. This can be a far more rewarding way to segment readers or an audience. In our experience, these clusters are a much stronger predictor of – for example, whether a user is likely to take part in a reward promotion – and the relevance of different areas of content – than standard segmentation based on value.
‘Natural segmentation’
It’s now possible to combine this kind of data with highly interconnected datasets that reveal the relationships between millions of topics. Unique algorithms can be applied to identify underlying patterns and trends that can be used to characterise the whole dataset. Clusters of interests begin to appear naturally and, using models that enable detect convergence to happen effectively, the most valuable targets in naturally occurring segments can be identified.
This is called natural segmentation because of the organic way the segments form or manifest without human influence. As data is based on a combination of ‘needs’ and ‘wants’ as well as actions, it is possible to develop a more complete and ‘human-like’ view of the audience. It’s a dynamic model, which is continually refined by new information from both behavioural and social sources. In other words, unlike traditional segmentation, it adapts constantly to change.
Natural segmentation has significant implications and benefits for publishers, helping inform editorial direction as well as enabling them to identify more accurate and effective advertising/audience segments.
This leads to better-targeted and effective advertising, in turn resulting in higher CPMs. The data can be plugged into a programmatic strategy or can be used for direct sales to enable marketing solutions based on rich audience data rather than purely content driven conversations.
So how can publishers put this blend of first party data and consumer insights at the core of the business? Here are a few sensible first steps:
- Establish a social login and create a strong call to action;
- Make on-site recommendations to readers, helping people to find and enjoy more content – with the associated benefits of increased dwell time, page views, and more effective advertising placements;
- Use aggregate data to make informed choices about strategy such as editorial direction for the next quarter, or where and how to target native advertising;
- Consider the value of natural segmentation, letting the data tell the story, rather than making arbitrary decisions on your audience, based only on, for example, on content, or demographics.