SAN FRANCISCO/LONDON: Automation of ETL (Extract, Transform and Load) promises to transform the lives of marketers by taking data used in campaign advertising analytics from multiple sources and combining this into a single database view for better decision making.

In a WARC Best Practice paper, David L. Smith, founder of media technology consultancy Mediasmith, explains how ETL gives marketers the ability to extract quality data from various sources, then transform that data so that it is normalized and can be translated into other formats and then to load it so that it is stored on an accessible system or database, one that is customized for marketers’ needs.

Given the volume of data now available – which can result in analysts spending more time simply handling it than actually delivering useful insights – ETL has the potential, he says, to bring efficiency and functionality into the optimization of digital campaigns.

Arriving at the end solution is not necessarily an easy process, however. Marketers must first identify which databases will aid better decision making and then determine how the data contained therein can be accessed and transferred.

That may require specialists who are able to confirm the veracity of the data and undertake such tasks as normalizing field headers and data formats. A suitable cloud service provider will also have to be chosen and reporting mechanisms selected which are appropriate to the viewing audience, whether C-suite, marketing manager or data analyst.

So far, so straightforward, but Smith highlights the likely internal challenge of a lack of data input standards, quality and consistency, which will necessitate centralized operations using an agreed-on dictionary and standard naming conventions.

And, he adds, “there remains a lack of consistency in sell side data” where there is a broad variety of inputs including from publishers, third-party ad servers, demand-side platforms, verification providers, social media providers, survey vendor and traditional media buys.

An example of how ETL can work comes from the Financial Times, whose Business Intelligence team worked with data warehouse Amazon Redshift to enable the publication to analyze its data using standard SQL and its own tools.

Redshift uploaded two years’ worth of FT.com behavioral data and added a reporting tool, for an initial load of approximately two billion records that covered every click on FT.com. This has given the publication access to centralized and real-time data to better inform the decision-making process.

Sourced from WARC