Marketing & Media trends

Marketing & Media trends

Retail trends

More Articles

#BizTrends2018: The near future of data fusion

The holy grail of media research will become reality in 2018, with the fusion of multiple hub and donor studies. Data fusion involves integrating multiple data sources to produce more consistent, accurate, and useful information than that provided by any individual data source. It is a statistical technique that creates a single data set from the attributes of two separate data sets.
We have already put 'fusion hooks' or relevant common variables on the Establishment Survey (ES), Publisher Audience Measurement Survey (Pams) and many other currency and industry studies.

The days of one-dimensional numbers are long gone and in media planning and buying, the ideal has always been to know as much as possible about the consumer; to enable more accurate and timeous branded messages. These include demographics, geography, mood, activities at time of impact, touch points during the day, retail behaviour, brand repertoire, media consumption and much more.

Much of this information is available, but it is scattered all over the place, contained within multiple studies, panels and sales data.

(c) racorn -

Single source, as far as I am concerned is like using one eye instead of two. Only with two eyes can we gain depth and perspective. It’s like using an outdated road map, versus step-by-step directions on Google Maps.

Upmarket consumers are not prepared to give 60 minutes of their time for an old-fashioned AMPS-type interview. As the world becomes more complex, markets fragment and data increases. Inevitably single -source data or research will no longer be viable as there are fewer respondents per survey question, and more questions are needed to understand the complexity of life and brands, which means longer questionnaires and poor quality of responses due to respondent refusal and fatigue. Besides which, single-source research is extremely expensive to conduct.

Fusion provides the key to consolidating all these insights into one easily accessible survey.

By fusing quarterly sales data from scanner panels, we get actual, recent purchases, not claimed usage. This allows for seasonal or even weekday media planning based on sales. This integrated data can also be used for post-campaign evaluation to measure media performance and ROI based on sales linked to exposure!

After just having returned from the Publishing and Data Research Forum in Madrid, I can state that of over 20 countries represented at the conference, there are only a handful (SA being one) that are not using data fusion. It is unquestionably and overwhelmingly the most popular media research methodology.

I am very excited for 2018 as we get vastly more insightful, recent and real behavioural purchase data linked to our media currency. This is the future of research and it is the fundamental model that Pams is built upon.

For additional information and reading research visit
Get a daily news update via WhatsApp or sign up to our newsletters.

About Peter Langschmidt

Peter has spent over 35 years in research, advertising, media and marketing. He runs a customer intimacy and research consultancy called Echo that specialises in work with media owners including Naspers, Kagiso, Caxton, Facebook, the Publisher Research Council (PRC) and the Interactive Advertising Bureau (IAB).