“Data Driven Thinking” is written by members of the media community and contains fresh ideas on the digital revolution in media.
Today’s column is written by Khurrum Malik, CMO of eXelate.
Water, water, every where,
And all the boards did shrink;
Water, water, every where,
Nor any drop to drink.
-- excerpt from “The Rime of the Ancient Mariner” by Samuel Taylor Coleridge
For many, “The Rime of the Ancient Mariner” brings back memories of high school or college. But one of the poem’s most famous lines -- “Water, water, every where, Nor any drop to drink” -- could easily describe the state of data today.
Data volumes and types are growing at a rapid rate, both in ad tech and other industries. It’s no wonder that terms such as public cloud, private cloud, data platforms, data management, data aggregation and, of course, big data are starting to meld together into a confusing mess. If Coleridge was around these days, he might be tempted to rewrite the key line to, “Data, data, every where, Nor any byte to think.”
In the ad-tech industry alone, marketers are drowning in the wide expanse of real-time data available, unable to wield it effectively to drive business value or improve KPIs. There’s first-party data, second-party data and third-party data. Can X-party data be far behind? What about PC data, mobile data, TV data?
Instead of creating an infinite number of names for data, it’s time to get behind some fundamental considerations to empower marketers and agencies to benefit from all the insights that data can provide. Before spending any time or money for more data -- or to act on the data you have – I’d recommend considering these four important factors: source, quality, scale and variety.
Source: From Where Does The Data Come?
This core question is a good place to start. As a brand leader or an agency representing a brand, you want to know that the data you’re about to use isn’t coming from an uncurated data source, often associated with social widgets, which frequently end up on uncontrolled long-tail publishers with brand-sensitive content. Instead, look for data providers who are investing in curated data sources, who are committed to both business and technical checkpoints to ensure data is coming in from brand-safe and credible sources. Examples of curated environments include panels (ex: comScore), social-media resellers (ex: DataSift), publishers (ex: Business Insider) and data providers (ex: Experian).
Quality: Is The Data Accurate?
Data accuracy is a sensitive topic that often generates contentious debate. When evaluating data, one should always ask not only how accurate the data is, but also if providers have set up an ongoing process to maintain data accuracy. Look for evidence of an ongoing commitment to accuracy by asking for examples of data validation from third parties and by making sure that accuracy isn’t being confused with precision.
It’s possible to be very precise and still be somewhat, or even totally, inaccurate. To take a simple case, let’s consider the target analogy. In this analogy, a measurement is analogous to a dart thrown at a target, with multiple shots representing repeated measurements. When throwing darts at a target, accuracy is determined by how close to the bull’s-eye a dart falls. Precision is determined by how tightly the darts are grouped together, regardless of where they are on the board.
Once you’ve decided to use a data set, don’t stop evaluating. Make sure you have a process in place to measure effectiveness and provide feedback to the source. Nielsen’s Online Campaign Ratings and comScore’s validated Campaign Essentials are examples of third-party tools that can help evaluate campaigns and the effectiveness of their associated data.
Scale: Is There Enough Data To Make An Impact?
In ad tech, more than in many other industries, having enough scale to drive a measurable business impact is key. Does the data against which you’re executing a campaign have the size to make a meaningful business difference? Even if you have great brand-safe and high-quality data, you’ll find yourself limited if you don’t have enough of it.
Ask your data provider what its data volumes have been in the past and present, as well as what volumes it expects in the future -- especially in the key segments you’re targeting. Acxiom is an example of a company that’s transparent about the scale and impact it can bring to a data solution. It references its list-rental database covering more than 126 million households and 190 million individuals.
Variety: What Types Of Data Are Available?
If you’re comfortable with your provider’s data source(s), quality and scale, it’s time to ask about the variety of data that’s available. Can it provide, for example, offline, online, personal, anonymous, mobile, PC, branded, blended, demo, interest and/or intent data? These are just a few of the types of data that come to mind for audience targeting.
Ask your data provider to talk you through its data taxonomy and its ability to maintain order. Be cautious of data providers that claim to be able to segment everything. Instead, ask for evidence of past work. For example, AirSage clearly articulates the range of data it includes in its population analytics suite.
‘The Rime’ Rewritten
As we move into an environment of the Internet of things and data ubiquity, keep these four considerations in mind when consuming and using data. If we can make these factors standard, perhaps we can all move to a time where Coleridge’s poem could be tweaked to:
Data, data, every where,
And all the issues did shrink;
Data, data, every where,
For us all to sync.