The stunning achievements and advancements made on behalf of humanity within mathematics are paramount. After all, it is thanks to engineers, mathematicians and statisticians that we ever put a man on the moon, that we've ever mapped the floor of the ocean, that human eyes have ever been able to see out-of-this-world phenomena the likes of colliding galaxies or the rings of Jupiter. In fact, it's arguable that mathematics is the very foundation of our physical world. The Fibonacci sequence, while prone to fodder for conspiracy theorists, can be found in every corner of the universe, from spiral galaxies, to sea shells, to the ratios of your facial features.
This omnipresence of mathematical influence puts a special kind of social importance on those working within the field. They are an elite class, the leaders of our Imperial Present, if you will. And with the rise of big data, it is again engineers, mathematicians and statisticians leading the way, acting as the stewards of permitting intelligence from the big data soup.
And though this computational elitism has achieved at a scale often bar none, the new field of data science needs to be reserved for the few – for those seeking to democratize the use of big data computation for the masses.
Our Collective Data Science Duty
Here's the thing, technology is empowering the public in never before seen ways, and data is the backbone of that shift. Between wearable tech and digital identity platforms, people are creating more data every day than has ever been created in decades, no, centuries past. Each of us is essentially our own personal data scientist, and those working in the digital space have very much been their own statisticians for quite some time. It's why platforms like Google Analytics, Omniture and more are so popular across the industry. They put the power of analytics in the hands of users, requiring little training but returning lots of measurability.
Big data platforms must do the same, providing for users the insights necessary without too high of a time spend. Sure, you can hire a data scientist to collect and analyze your user data – after all, this is exactly what the New York Times did – but the cost is high and the return only as relevant as the accuracy of their predictions. Besides, there is a growing number of college educated, yet underemployed workforce out there, people who have been trained on digital analytics by their own accord (i.e. from their personal usage). This new talent means fresh brains, fresh eyes and, ultimately, fresh ideas.
"Each of us is our own personal data scientist, and those in the digital space have very much been their own statisticians for quite some time."
Here, I break down how to approach big data with a data scientist’s mindset, and get even better results.
Stepping Up to the Big Data Plate
Analyzing big data is as simple as selecting which direction to go when you come across a fork in the road. You have two choices – either you want to know a past result or a future likelihood. There is no such thing as real-time data. By the time you get it and analyze it, you are already searching for a past result or a future likelihood.
Big data, then, only has two approaches, and your team must choose which approach to use for each big data analysis.
A common misunderstanding during data analysis as it relates to marketing and digital analytics is not considering the actor of the past or future behavior, essentially the "who" behind the action. Keep in mind that the definition of who can be a single individual or a segment of individuals that all share a trait (i.e. viewed the same web page). That said, data is not merely the collection of billions of 1s and 0s. It is, instead, the collection of individual interactions with digital entities, whether that be a Facebook like, a pageview or what have you. There are real people at the genesis of data and in order to put profitable action to big data, it's critical to remember that it is generated at an individual level.
Make Big Data Small, Then Scale
Break your data into segments, and target messaging, strategy and the like in order to please the end-user. Figure out whether you want to know a past result or a future likelihood, and then get granular.
In the end, big data needs to be made small in order to produce ROI. All too often, present thinking surrounding discovery into big data and analytics seems to be a simple equation of adding more data scientists and more algorithms – but that leaves the responsibility of data valuation to too few.
Increasingly, we find that while we are told we live in a super smart new information age of real-time digital efficiency, there aren't many access points. Individual teams within organizations – often the sales teams, marketing teams, account managers and executives – want answers from their company's data in order to provide improved guidance and eliminate extensive guess work that may or may not have any real payoff.
"Democratize big data so all departments within an organization can discover and act upon customer adjacencies."
Democratizing big data so that all departments within an organization can discover and act upon unknown customer adjacencies or affinities is how to unlock the elusive digital efficiency. Umbel is how you offer that democratization.