As a natural resource, crude oil is full of potential value. Yet without the refining process, this energy powerhouse is little more than a bubbly, brown liquid.
Increasingly, data is replacing crude oil
as the king of resources. This seems like great news, as businesses are collecting so much data every year that we're practically swimming in it.
The paradox is that, while an increase in data volume gives companies more to analyze, most of them are struggling with how to make use of it. It's as if organizations are sitting on an enormous oil reserve, but lack the infrastructure -- the refineries, pipelines, and personnel -- to turn the crude into all sorts of useful, interesting things. They are data-rich but insight-poor.
Despite the struggles, it is possible to leverage data science and machine learning to scale your business, save time and grow revenue while improving your customers' experience. I believe the key to unlocking the full value and potential of data lies in two places: the data itself and the people employed to make use of it.
Take Stock Of Your Data
When getting started with data science, begin by considering how your organization views data: Is it seen passively, as a byproduct of transactions? Or, in the vein of businesses like Amazon, Google or Uber -- do you view data proactively, as a strategic asset?
Next, look at the data you have and ask what else you'd like. Transactional data is just the beginning. One of the most common and critical types of business intelligence is customer data. Within that realm are categories such as contact data, customer preference data and customer telemetry data (i.e., what they're doing with your product). Each of these data types come with their own opportunities and challenges, but no matter how you classify them, collecting customer data across various touch points will set you up to deliver rich insights.
Finally, decide on a data acquisition strategy. Identify the types you want, figure out how you're going to gather it, and determine metrics to track progress toward your goal, whether that's improving the amount, variety or quality of data.
You can gather customer data, for example, by asking customers for it with a new web field or requiring them to register for content. You can purchase/lease data through third-party groups or share it with partners. There is often a lot of data already existing in a company that's simply unconnected or siloed in different databases
Consider Investing In Data Scientists
The types of data alone are overwhelming. And the reality is that the Excel ninjas and dashboard experts within your organization -- with skills that worked so well 10 years ago -- can only take you so far. You need to predict what's going to happen so you can take the right action. And the key to unlocking those future-warranted insights are data scientists.
Bringing in specialized tools and vendors can help, but it's critical for you to have your own capability: Evolve your traditional data analysts into a data science skill set or hire data scientists with expertise working with bigger data sets. The right data scientists can deliver a huge amount of value in a very short period of time. One data scientist on my team recently solved a problem -- in his first five months on the job -- that will have a multimillion-dollar impact over the course of the upcoming year.
The Right Data + The Right People = Business Impact
Practices such as machine learning, predictive analytics, and data science modeling are a bit different from each other, but they work in similar ways. In each, you start with a problem to solve, use data scientists to find data they can use an algorithm to predict what's going to happen and use the results to improve your prediction for next time. That prediction can change and improve and deliver incredibly powerful insights.
In the healthcare industry, machine learning is helping to connect and reconcile patient records from across silos much faster and more accurately than before. And in the same way that Uber uses data to identify peak demand times for surge pricing, you can use a similar concept turned on its head. For instance, think about evaluating interaction data with customers to determine what their "demand for engagement" is.
By analyzing data from across contacts, you can see patterns in terms of their likelihood to engage, allowing you to target your outreach in a very personalized manner for every contact you have, maximizing the likelihood of an opened email, click, like or share.
Understanding patterns and trends are important. But, armed with better data and skilled data scientists, you can do much, much better -- you can start to unlock the true value of data science and maximize the insights you can gain from it.
The article was originally published here
As director of data science for Cisco since 2014, Sanjiv plays a leadership role in the company- and partner-wide digital transformation,