At a hotel close to the statue of Columbus in New York, Oracle's general manager of its Data Cloud, Omar Tawakol, explained how he has set his sights on helping data-as-a-service (DaaS) users find a new world of data quality. But it is not the traditional, technical data quality you might expect; this was about the next step – ensuring that business consumers can derive actionable insight and build confidence from the use of DaaS assets.
Confidence is the new quality
In his keynote, Mr. Tawakol argued that data is an "invisible asset," and that the incentive to "water it down" to either sell more or grow marketing reach is strong, often at the expense of data quality. I agree. The once-popular claim that data is the new oil is an analogy that just keeps giving. Like oil, data must be carefully sourced, refined, and enriched to grow value, and then connected to the right consumers. The process is essential if DaaS is to deliver on the promise it holds; in cases where data assets are acquired and fail to deliver due to insufficient quality, the fast-paced world of digital marketing will soon move on.
But how do you assess the quality of DaaS assets? Mr. Tawakol offered four primary methods:
Scoring on quality – For example, is this person identified in the data likely a male or a female? Providing the user with a score of confidence is important here.
Cross-validation – The comparison of multiple sets of data about the same subjects. An appealing method which could also be automated and enhanced with emerging machine-learning technology.
Outcome measurement – The "holy grail," that is, the ability to identify that particular attributes drove a particular outcome, e.g. a purchase. The catch being that it is very hard to do in reality.
Scoring – Somewhat like a FICO credit score, understanding the propensity of a person to behave in a certain way has value. For example, what is the likelihood of that person buying a particular brand of car?
Each of these methods offers benefits. My view is that it will not be an either / or approach. In fact, given the ability to automate cross-validation, add a scoring mechanism, and enrich the whole process with machine learning to continuously improve the quality suggests that there is an obvious path to be taken. While outcome measurement may not always be possible, it will remain the ultimate – if hard to reach – goal. Whatever approach is adopted to grow confidence, DaaS vendors, like Oracle, have a central role to play; given their familiarity with the data they are providing, and significant technology capabilities, the opportunity to offer a growing portfolio of cleansed, enriched, and verifiable DaaS assets is substantial.
A concluding thought: Growing confidence in DaaS is important, sure, but why? Increasingly, the focus of digital marketing is on maximizing the value of each dollar spent, and targeting the next dollar as effectively as possible. That's why Oracle is talking about data quality in DaaS – digital marketing is maturing and rather than being a poorly understood "me too" spending splurge, is ever more subject to the usual enterprise expectations of return on investment.
Tom Pringle, Head of Applications Research