R2i's Co-Founder, Chris Chodnicki can often be heard talking about the "last millisecond" when it comes to discussions around data. And if you're in marketing, or analytics, or any related field, you're hearing a lot about big data. However, as was shared on stage at the Inbound Marketing Summit this fall in Boston, big data is really all about small data. Bob Collins, Content and Digital Strategy Director of the Pulse Network interviewed Chris as well as r2i's VP of Digital Marketing, Eric Jones to get their take on data, big and small, what it means to marketers and how to use data insights to understand KPIs and actionable takeaways from dashboard reports. Here's a recap of their fireside chat.
What do marketers need to know about big data? How do you define big data?
Big data is a volume of data. It could be structured or unstructured and it doesn’t come from just one source. Social media for example has contributed to an explosion in data sources. With so many data sources, organizations rely on data warehouses that can store tera and petabytes of data.
Big data is also a result of frequent data collection which carries some velocity with it that can quickly overwhelm a marketer. Think of the amount of data that can be pulled from any given social stream, real-time activity or geo-sensors. In this way, big data carries a lot of emotion with it and doesn’t tell the whole story.
This variety of data sources adds up to multiple disparate data sources from everything from transaction data to data pulled from owned and un-owned properties.
In order to make use of big data you need insights to create data intersections and segmentations. You need to insert a human being into the data analysis process.
Why is relying on big data alone shortsighted for marketers?
There are too many risks in relying on big data alone. It can give an emotional view of a story but big data cannot tell a whole story without human insight there to interpret the meaningful data points that are worth exploring, slicing and segmenting further. Because working with big data cannot be a fully automated process, you have to factor in certain areas for risk such as human error, invalid, inaccurate and inconsistent data. There is no black box that can accomplish the human’s job of dealing with large volumes of data to find the relevant intersections.
We come back to the need for human insight which helps to create processes, identify KPIs and set-up the measurements against those KPIs. It’s the human who handles the data interpretation.
How do you define small data?
Small data sets are created when humans get involved to create and interpret data relationships formed out of multiple data sources. You might filter by different data slices like timeframes, geography or profiles to create specific data segmentations. You watch for outliers, trends and out-of-context data. Small data lets marketers meet their customer at the last millisecond of their activity. Having the ability to make better business decisions based on the information small data provides means that marketers have access to insights that can effect a customer’s decision at the right moment.
What are some of the most valuable insights a marketer can get out of small data sets?
Take a social analytics dashboard as an example. Big data might be the reporting view of just one social channel but when we overlay and compare with multiple social channels we can identify key areas of success and opportunities for optimization. Small data informs us what content is working, what channels are most successful, what timeframes get the most engagement. We can then overlay the social data with other metrics for a more refined view of digital marketing. By looking at small data sets we can pinpoint realistic KPIs. We can also drill down into the exact content types, themes and messaging and evaluate if we should continue down a particular path based on results, i.e., more video, more infographics, etc.
What are some key takeaways marketers can use to get started in leveraging small data?
Identify all the data sources you have—site analytics, social streams, transaction information, un-owned property data.
Establish your KPIs and the metrics you need to capture. You can use data intersections and cross-reference different data sources to confirm your KPIs. You can also identify many of these key data points in an aggregate manner. This will also help to identify specific metrics inside each stage of your customer funnel and identify things like transactional precursors, and of course conversion points themselves.
You will also need to define your targets and goals. How do you want to segment your audience? What slices of data will need to be analyzed to help identify these segments?
Make sure your marketing team and your IT team sit next to each other. You will need to leverage technical tools for aggregation and reporting purposes.
Then you test and adjust. The power of small data is that it tells you where to optimize. You can analyze trends and outliers using analytics and visual data tools. Learn what works and what doesn’t so that you can refine until you are intercepting your customer at their final point of decision-making.