My answers have varied depending on the person asking and the context of the question. And while I take pride in my ability to distill business benefits from highly technical topics, I’m pretty sure my answers weren’t that helpful ... until I read The Gap Between Big Data and Big Insights : Turning data into engaging stories from Brian Solis.
Now, I feel pretty confident that I can address the big data question with grace. A few lengthy discussions with clients have fully tested my viewpoints and in the process, I’ve actually developed a clear understanding of the useful nuggets central to this overhyped topic.
“The problem with big data is we think that by saying ‘big’, we automatically convey importance and urgency up, down, and across our organization.”
As Mr. Solis points out, this simply is not true. And to be clear, big data doesn’t have to be voluminous either; it is “big” if it represents comparatively large data sets that may be analyzed computationally to reveal patterns, trends, or associations, and especially where the data points relate to human behavior and interactions.
What follows in this article is the story of how I used extremely tiny slices of big data to gain insights concerning call log data for a very small company. It demonstrates that big data opportunities are increasingly within reach for even budget-minded businesses that want to use data and automation services to become more innovative, competitive, and operationally efficient.
Case In Point
Imagine you can harvest every call event going through your virtual PBX. Further imagine an automated process that maps every inbound call to your reservations extensions (your company is a transportation provider). Using integrated reverse-lookup services you are able to determine the approximate location of the caller.
And using the location of the caller, it's pretty easy to map the weather in the caller’s area.
Now you’re a big data user.
You’ve created a mashup of real-time inbound calls, locations, and the general weather conditions that the caller is experiencing. While your reservation agent might appreciate such timely and relevant information during the conversation with the customer, there are even greater opportunities to leverage this information to create insightful awareness of future customer patterns.
In the case narrative I mentioned the reverse-telephone lookup service. What if each lookup event was also considered a valuable data point?
Imagine if we captured analytics about the lookups themselves. Now we have more data - subtle [additional] insights cached and counted and all in the context of our original PBX data. We now know that when a call comes in from area code 970 in exchange 928 that it's a rare occurrence - an outlier.
We tend to assume that Google Analytics is for web activity alone. However, Google Analytics provides far greater opportunities to blend physical devices and other virtual elements with events and location over time. Indeed, nearly anything can be tracked and analyzed in Google Analytics by using its simple submission API.
Imagine we push all of the PBX call event data into Google Analytics; we can blend and align web landing page activities with PBX reservations events.
The CEOs Dashboard
Along the pathway to creating insightful indicators, we can leverage the data to show calls, call minutes, and reservation minutes for the recent week.
Nightly Call Summary Report
At 10 pm every night, my client’s management team receives a synopsis of the day’s call activities. It is an infographic-style report conveyed through email and contains various metrics about the day’s activities.
By assembling various statistics and computations combined with locality, the report provides daily insights concerning the most recent 22 hours of call center activity.
Given time, location, and call duration, we can learn more through GiS mapping techniques.
This example “torque” map allows us to see where calls are originating and when. By color enhancing the calls and using call duration to subtly adjust the size of the points, a deeper understanding of call center behaviors begin to take shape.
Since each call and resulting location lookup event is associated with time, we’re in a good position to make predictions about callers in the future. We can also determine if certain weather conditions inspire people to call into our reservations system.
Given these simple data intersections and observations, we can predict with relatively good success, that the call center will receive a predominant number of reservation inquiries from the central eastern coastal states between 4 and 5 pm MDT when snow is falling in or near the caller’s city and where all other parts of the country are sunny, or mostly sunny.
Armed with simple analytical models derived from these mashups, we can use weather forecasts to anticipate likely call patterns, making it possible to prepare workers for certain types of itinerary demands.
This is a simple example using big data to create small insights.
Big Insights, Small Cost
The case narrative described above is not a demo; it’s a fully automated system that’s in production and it was entirely built with Google Apps. It is a lights-out operation - no humans need to export and import raw call log data, and none of it requires any attention for the data to be utilized across the company.
This approach puts big data, small insights into financial reach for small and medium-sized businesses.
The client, however, is convinced that these are not small insights; rather they’re sizeable ones given the impact and availability made possible through automation and Google Apps. Apps is a powerful and worthy data environment that is inexpensive and comfortably utilized by the firm’s workers.
Spreadsheets replace SQL databases; heavy-handed code is replaced by lightweight scripting services that are easily automated with script triggers instead of complex server-intensive cron programs.
Google Apps provides all of the necessary elements and services required to create and manage solutions that provide access to large and small data sets.