By a show of hands, how many people have heard the term “Big Data”? If your hand is not in the air then you are either lying, asleep in every single meeting or live under a rock. In all seriousness, big data has been the topic du jour in the marketing space for the last five years. But what does it really mean? How do we provide value to our clients in this space? What role do we play as the agency? Do we simply take every report and dump it into a magical, mystical tool built by Oracle, Teradata or Microsoft and then poof! Marketing insight!

I wish I could say this was the case. Designing a cross-channel measurement strategy is really as many parts art as it is science. A former boss of mine called it the “Art of Measurement and Science of Insight.” Sounds a bit backwards, but it is true. For most enterprise marketing organizations, the art of preparing to create a big data strategy is a bigger hurdle then actually analyzing the data itself. As an agency, we have a great opportunity to help.

Many marketing organizations have a hard time understanding what their full data footprint looks like. This problem is usually due to their team structures. Generally, there are a number of departments handling touch points throughout the customer life cycle, each with one to many roles managing one to many channels. When we consider that each role is focused on getting its communications to market in an efficient and timely manner, we see that there is little time for each to gain broader understanding of where the rest of the marketing organization’s goals and data reside. Agencies are actually well positioned to help discover and map out an enterprise’s marketing data footprint. As third parties sitting outside of the organization, we have a cleaner line of site into all marketing programs, stakeholders, vendors and data stores across the enterprise.

Another opportunity to assist in our clients’ big data strategy involves the issue of varying levels of data granularity. I know, this is a very scary, nerdy term. It simply means that data we generally pull is in the form of reports. The problem with reports is that they are rolled up or aggregated views of information. If two or more reports are rolled up in different ways — whether it is by time, region, campaign, etc. — and then combined, the data won’t make sense. We cannot create a cross-channel view of campaigns and customers. The key to success is to gather the most granular data possible. Most of the time this data is down at a customer behavior level. At the customer level, we have the opportunity to create a picture of the consumer across all channels. Then the data we retrieve can be rolled up to create the aggregated views mentioned earlier and many more. We as an agency have the opportunity to work with vendors on our clients’ behalf to request and gather more granular data on a regular basis.

The most common item overlooked in big data solutions for marketing is context. Big data solutions such as Teradata Aster and many others are designed to identify correlations in behavior across millions of pieces of data. Very few system-generated data feeds account for marketing context. This is the data that is generally included in the marketing brief — budget, regions, offers, products, segments, price points, channels, creative, message tone, etc. Analyzing the correlation among marketing data points is basically meaningless without this context by which we query and slice and dice information. Agencies are the key creators, managers and contributors of this information, yet we rarely hand it over, let alone analyze it. If this information is collected, tagged and organized in the proper fashion, it can be an extremely powerful tool on the path to marketing insight when combined with behavioral and campaign data.

With the world shifting to an always on, connected environment, there is no hiding from big data. Marketers and agencies alike are scrambling to figure out what it all means. The key to a successful solution and happy client is in the artful collection and organization of the right data, not the big fancy tools or buzzwords.