The right (and wrong) way to make data-buying decisions

There are major industry forces in play that are changing the dynamics of how data gets generated, collected, and used for optimizing customer interactions across offline and digital touchpoints. Digital data expansion has accelerated the "big data revolution." In fact, 90 percent of the data that exists in the world has been created in the past two years and includes valuable anonymous data sources in addition to personal data.

According to industry sources, digital media spend is expected to grow nearly 50 percent and overtake TV ad spend by 2016. Targeted display alone will grow more than two-fold by 2016 and surpass search as the No. 1 digital media category. Data-informed programmatic media buying continues to grow as a way to generate high performance from digital media.

The right (and wrong) way to make data-buying decisions

In today's fast-moving digital ecosystem, there is a lot of hype about digital data, but its value is not fully understood, and its accuracy not properly validated. Leading digital data players build impressive technology stacks and do a nice job of collecting and aggregating various digital data sources, but there continues to be a major gap in analytics.

Existing audience selection practices for digital targeting are primarily driven by off-the-shelf, generic third-party segments that are not optimized to specific campaign objectives. Marketers are not making use of rich digital and transactional first-party data to create custom audiences and are not maximizing the value of data that is available to them. We see some digital marketers with limited analytics experience making data-buying decisions based on subjective factors such as relationships and market hype.

Some digital marketers gave up on the idea of using third-party digital data completely and believe they are not getting enough benefit from the use of third-party data to justify the cost. On the other hand, data sellers arbitrarily set the price for their data assets without any consideration into the incremental lift it generates for digital campaigns and wonder why digital marketers don't leverage their data to make targeting decisions. Others, who make data-buying decisions based on subjective factors, get disappointed by the accuracy of the data when they realize that a big portion of the female audience segment they purchased in fact consists of male consumers.

Transformation of the digital data landscape

Digital data sourcing needs to undergo a major transformation. To put this transformation into context, we should take a look at the changes that have taken place in the data ecosystem over the last 10-plus years. In the early 2000s, content-based buying was used as a proxy to reach desired audiences. In the mid 2000s, we saw continued reliance on content as a proxy, but re-marketing provided a shift in targeting. In today's environment, spend shifted toward third-party data, and first-party audience targeting grabbed the interest of marketers. The future state is all about using custom predictive analytics driving media targeting decisions and integrating offline and digital data sources. Digital marketers who embrace this transformation will have a tremendous opportunity to generate high ROI through analytically led digital data solutions.

What digital marketers need

The transformation of digital data sourcing and targeting needs to be driven by several factors. Measuring the value of digital data sourcing requires advanced statistical techniques.

Unbiased approach: Analytic evaluation and validation of all available digital data using an agnostic approach is key to determining the most valuable data sources.

Big data capabilities: An internet scale, real-time big data platform is needed to process vast amounts of digital data and integrate with DSPs, ad exchanges, and other digital media partners in the ecosystem.

Advanced analytics expertise: Predictive analytics and optimization will enable the mining of granular first-party data such as response, conversion, and customer value to develop custom audiences for digital targeting.

Transparency and privacy: It's important that marketers leverage all the available data using privacy guidelines and provide full transparency to consumers in terms of how their data gets collected and used to create relevant messages and offers.

The output of this analytic evaluation process is a detailed, customized scorecard that marketers can use to rank order various digital data sources that are available in the market place. This is a much more rigorous method than single-dimensional simplistic methods like lift reports that are widely used in the industry. This type of analytic evaluation and ranking of data sources leads to optimized audience targeting practice and higher campaign performance.

Marketers need ask themselves some tough questions when assessing a source through an analytic evaluation process. They include:

  • Does the source add incremental lift to my predictions? This involves the evaluation of third-party data on various models such as engagement models, response models, conversion models, and customer value models.
  • Does the source provide accurate and high-quality data? This involves matching responder data against high-quality, multi-sourced variables to certify the accuracy of key attributes such as income, age, and wealth segments from the other sources.
  • Does the source provide access to new and unique prospects? This involves gauging the overlap and understanding incremental unique coverage from each provider.
  • Does the new source provide the ability to better segment my target audience or lend new insights? This dimension tells us how a particular data source improves our ability understand and segment customers.

Conclusion

The digital data landscape is changing rapidly due to innovation and the big data explosion. An analytically led, unbiased approach is needed to identify the most valuable digital data sources that will drive high performance for marketers. We are in the early stages of an exciting journey and new approaches will be necessary. First movers embracing advanced analytical methods will have a significant advantage.

Ozgur Dogan is general manager of the data solutions group at Merkle.

On Twitter? Follow iMedia Connection at @iMediaTweet.

"Puzzle and binary code" image via Shutterstock.

 

Comments