How Small Businesses Can Pull The Right Data From Big Data


At the moment you’re reading this, there is more data than there ever has been before in human history.

In fact, if you re-read the above sentence a thousand times, it would prove true every time. That’s because the amount of data produced by people per minute is staggering — 2,657,700 gigabytes of Internet data every minute in the US alone, according to an article published in July 2017. That’s just one slice of the pie, and as our population and technology use grows, so will the amount of data we produce.

This data explosion is due in large part to the growth of Internet of Things (IoT), which includes smart refrigerators, televisions, and all other sorts of internet-connected electronics. These sensors all produce information of some kind or another, and most are connected to data warehouses, contributing to the ever-growing sea of Big Data.

On one hand, the exponential proliferation of data means that small and large businesses alike can better improve the customer experience and identify new opportunities by applying the right analytical measures. On the other hand, the sheer amount of information that these companies have to sort through can make it hard to know which data is quantifiable and which data is just… well, data.

Big Data Basics: Use Caution

First you’ll want to look at where you’re acquiring your data from, and how you’re processing it. Just because you can pull data from anywhere nowadays doesn’t necessarily mean that you should, or that this data is going to be accurate.

For example, Villanova’s School of Business mentions that social media is a great way to help capture quantifiable data. They explain that if somebody (a private individual) in a brand’s social community announces an engagement, a businesses should recognize and capitalize on the marketing opportunities that follow, including flights, hotels, travel deals, and wedding and shower gifts.

However, this initiative must be executed carefully. In 2014, Pinterest made the mistake of sending emails to a number of users congratulating them on their upcoming weddings — even though they weren’t getting married. Because the data showed that these users were highly active on wedding boards, it was incorrectly assumed that they’d be getting married, leading to embarrassed and even angry customers. Yikes!

This is a pervasive problem with Big Data — just because you have a ton of data doesn’t mean that it’s accurate, or that you’re using it correctly. The consequences of incorrect data analysis might seem trivial in some industries, but when it comes to big data application in predictive policing, for example, the consequences could mean the difference between freedom and incarceration. Tread lightly, and check, double check, and triple check your analytics before taking action on said data.

Selecting the Right Data

Maxwell Wessel, writing for Harvard Business Review, suggests that to keep an innovative edge on competitors and suss out the right data efficiently, business owners and data scientists need to ask the right questions:


  • What decisions drive waste in your business? Having a target in mind before you even shoot for the data is a great way to begin acquiring and analyzing it. The example used by Wessel is the floral industry. “The average retail florist can sustain spoilage rates of more than 50% of their inventory,” he writes. “More than half of their flowers simply become refuse.” By focusing on finding data to reduce spoilage rates, such as how many flowers to order, grow, and stock,  innovative flower companies have been able to reduce waste immensely and become disruptive forces in their industry. It all starts with knowing what you want to do.
  • Which decisions could you automate to reduce waste? Wessel argues that once you’ve figured out which decisions drive waste, the next step is figuring out which of these decisions can actually be changed. “Simple, repetitive, operational decisions,” according to Wessel, are better left to machines. In the case of the floral industry, this might simply come down to how many flowers to order, but another example he uses is the pricing strategy of America’s largest internet retailer. Amazon has apparently done away with a human pricing team in favor of mostly algorithmic pricing. “For most retailers this would be blasphemous,” he writes. “But if Amazon’s algorithm works, it would translate to far less spent on discounts, far less inventory piling up in warehouses, and better predictability of new product introductions — each of which would yield enormous competitive advantage”
  • What data would you need to accomplish all of this? After figuring out where waste exists in your model and what you could automate to reduce, it’s time to seek out the right data to make it all possible. In the running example of flower shops, that might simply come down to comparing inventory orders against customer purchases at different times of the year, during different marketing campaigns, and in relation to what customers are purchasing. If you found out that people don’t buy as many daisies in March, but that daisies sell like crazy during your April showers campaign, you could automate your orders and buy fewer daisies in March based on actionable data.


What Wessel has done essentially is flipped the common way of thinking about all of this on its head. Instead of looking at a giant data set, then trying to figure out what you can and can’t automate based off that data, and then hoping that you’ve made an efficient decision, you do almost the exact opposite.

A great analogy for the way that too many companies use big data is that they have all of these arrows, all of this ammunition, but all too often they fire first, and try to paint a target around where they hit later. Don’t fall into this trap. Select your ammunition wisely, and always have a target to fire at.

Related Posts



No comments made yet. Be the first to submit a comment
Already Registered? Login Here
Thursday, 12 December 2019
If you'd like to register, please fill in the username, password and name fields.

Member Login

Business Insights & Tips


Jill Holtz
2343 Points
Tena Glaser
1395 Points
Michael Lane
802 Points
Ron Immink
732 Points
Fionan Murray
721 Points
View Leaderboard