At The Northridge Group, we specialize in helping customers make sense of their data by using data analytics to solve problems with fact-based insights. Over the years, we’ve learned to take the mystery out of turning data into information. We’ve compiled a few steps that can guide you in translating your big data into big insights.
Identify the problem first.
Albert Einstein once said, “If I had an hour to solve a problem, I’d spend 55 minutes thinking about the problem and 5 minutes thinking about solutions.” Similarly, when faced with the equivalent of millions of file cabinets full of data, it’s critical to first identify what business question needs to be answered with this data. It may be that you’re looking to identify the root cause of a maintenance issue, or the reason behind a sales forecast shortfall, or the triggers behind an increasing cost trend, or to evaluate customer behaviors across multiple channels, such as the web, smartphone apps, social media, and the call center. And, sometimes, you may not know the specific question that needs to be answered but have a general sense of what needs to be examined. That’s OK too. When analyzing data, pinpointing the question to be answered or even just the target area to be explored, is the beginning of your “big data” journey.
Look beyond typical data sources.
Big data is just that, and because it’s big, there are no boundaries to the data sources that can be included in your exploration. Yet the sheer volume, variety and velocity, three “v’s” of big data, can be both daunting and exciting. To get the most out of data, look beyond your familiar departmental sources. If you’re in a customer service organization, there is a wealth of data in finance or in operations or the network or from your vendor systems. Think broadly when identifying the variety of data sources to include in your “big data” solution.
Normalize your data.
Compiling data from multiple sources requires a normalization step. For example, if you want to use data to solve a customer issue, you’ll find that not all data contains the same customer identifier. You will most likely need to connect the dots across data sources to create links between files. One file may contain a customer’s phone number, while another has an email address, and another a physical address, or account number or trouble ticket number. Don’t limit your data sources just because there is not a straightforward way to link data among files. Examine the data fields available in disparate data sources and build a map that allows multiple data sources to be linked together. There are tools, such as Hadoop, that facilitate the storage and processing of large data sets, including normalization.
Prepare for data mining.
Identify the metrics or facts you want to extract from the data. Think about the trends, patterns or anomalies that will provide answers to help solve your problem. Think about your data, which has now been assimilated and normalized, to identify the historical trends or patterns that will solve a current problem or predict future outcomes. For example, you may want to understand a customer’s experience as they move from an app to the web, to social media to the call center – the frequency, the timing, the intervals of their contacts. Coupled with that, you may want to understand if it was a sales or service issue driving the need for a contact. Furthermore, you might be interested in knowing if there is a time of day, day of the week or geographic pattern to the contacts. In this step, you’re identifying what to look for in the data.
Analyze your data.
Only after these first steps are completed are you ready to actually analyze the data. Performing data analytics takes more than the latest BI software or the brightest team of data scientists. Analyzing data requires people who have an understanding of the business and know what to look for in the data. To draw a parallel that is far from technical, if you’re searching for gold, you first need to know what a gold nugget looks like – it’s not enough to have the best digging tools! Much the same is true with big data analytics. Tools are important, but you also need talent who knows when they’ve found a golden nugget in that data! Business-savvy people with industry and operational expertise are key to making the most of your data.
Present insights in a meaningful way.
Finally, determine the most effective way to present the trends and insights discovered from your data analytics. Portray the results in a visual display that enables decision-makers to easily understand the value of the data and then to identify the necessary business actions based on those new insights.
Big data is a hot trend, sometimes clouded with some apprehension and uncertainty because of its complexity. However, with a few clear-cut steps, you can use big data to derive big insights.