October 5, 2017

Making Big Data Practical

 |  By: Mike Opperman

The best and worst part of using technology is the amount of data available to help with decision making. On one hand there is more data to provide insights into key performance indicators that can improve management efficiencies. However, picking the important aspects of the data avalanche can be hard.

Because, after all, the data is really useless if you can’t pull information from it that can help you run your business. “Big data is like a big trash dump. You have to know how to find the nuggets so it’s profitable,” says Vin Gupta, founder of and

“It is better to collect less information, but something more relevant, and then use it to improve your business.”

The key with big data is not to get overwhelmed yourself, or worse yet, overwhelm your employees. Choosing a few key reports can help you manage data effectively. “We provide dairies with five to seven reports that give producers the capability of running the dairy without overwhelming them,” says Nancy Charlton, DVM, dairy management adviser with DeLaval. She and her team, Chris Duffy and Andre Aguiar, Ph.D., advise mostly owners of robot dairies on how to manage data from those systems. Beyond those initial reports, Charlton says her team assesses a producer’s ability to handle more data as they become better at making management decisions from those reports.

Producers with large employee teams need to assess the ability of direct reports to understand and disseminate data. Overwhelming them with reports that aren’t critical to their job performance directly impacts the efficiency of the dairy. “In most cases problems are very simple, and the answers are simple, too,” Gupta says.

When employees feel overloaded, that’s when the basics get overlooked. “There’s so much we can do but I want to see basics done right, then we can talk about what to do to take things to the next level,” Charlton says. “If cows aren’t clean and comfortable, if there isn’t fresh feed and water available, then any kind of data won’t be able to provide you with the results you want.”

It sounds simple, but one way to make data more effective is to make sure reports are tied to protocols. For example, Charlton says a report that shows any cow at less than 60 days in milk that is less than 100% of yield from the previous day should be monitored. Cows are creatures of habit, so any deviation that can be determined through activity monitoring should be monitored. “If a cow does something every morning at 8 a.m., and she’s not doing that or hasn’t been doing it, then we need to ask why,” Charlton says.

However, she says in robot herds cows milk when they want, so it can be a hard to know when to take action. 

For weekly and monthly reports, understand the context from when the data was created. A snapshot that shows one point in time doesn’t provide meaningful information about your business. Instead of looking at one piece of data, try to identify trends that paint a larger picture. This can help you better understand the direction your business is going, and likewise any critical adjustments that need to be made.