The Human Element of Bad Data, Internally or with Your Agency

by Marketing Team,
Wednesday, November 29, 2017

In today’s data-driven world, having access to relevant and clean data is a powerful driver to success for brands worldwide. With the modern advancements in technology and data expertise, it may be surprising to know that bad data still plagues many of the world’s leading brands. Bad data is more than just an inconvenience though, it’s expensive. In 2016 alone, bad data cost leading brands over an estimated $3 trillion in the U.S.

There are numerous causes of bad data and they range from complex to quite simple. Working with top brands across the globe, one issue comes up more than most when we encounter bad data: human error. While human error is largely unavoidable to varying degrees, many of the causes of human error in data are, in fact, preventable.

Adjusting to Big Data

Although data has become more important than ever, essentially becoming a new currency, the treatment of data and those who work with it, has not adjusted adequately. When it comes to agencies, this is a prevalent issue as data is often neglected in the overall project structure. Data management is typically a low priority at an agency as they must focus on client deliverables or going to market. Furthermore, most brands don’t typically task their agencies with providing full data – so of course, agencies don’t focus on training and proper education on data collection.

Agencies are not the only data mis-handlers, most companies are not structured for excellent data management. Often times, data entry is manually handled by an assistant or other associate who is responsible for data entry at the most basic level. Data is typically not what that these employees are interested in, trained for, or equipped to handle correctly. These team members are not specialized in data handling and yet they are trusted with one of the most valuable commodities, data

If hiring a Data Manager is not an option, there are still steps to take to improve data management. Working with brands on their agency relationships and advising on their own governance and team structure, we provide the following basic steps to proper data management:


  • Invest in proper training. Any employee that works with data should be trained on your data technology, be it Excel or any other data management tools. Making sure that they possess expert knowledge and are familiar with best practices so that mistakes can be avoided.
  • Enact a data review process. For any company interested in becoming truly data-driven, there should be a stricter review process both internally and with an agency. Quite often, due to a supervisor’s workload, an assistant’s work will not be checked for accuracy; a second set of eyes can often find errors that were initially missed.
  • Don’t overlook data issues. When an issue with data is found, the typical course of action is to make a correction based on guesswork. This action can cause a compound effect of errors and create more work for the next person who needs to use the data. Instead, reach out to the data creator and investigate the root cause to see if a larger issue is to blame. 


Be a champion of your own data

Data isn’t going to lessen in importance so it would be in a brands best interest to begin investigating how their data is being handled both internally and within their agency. When it comes to the agency, inquiring about specifics such as team size, workload, training, and turnover rate can provide a good start to understanding their data management capabilities.

Internally, brands need to ensure they are set up to properly receive and analyze proprietary data. By setting up a governance framework and internal team training, brand teams can start utilizing quality data to make business decisions. No matter if internal teams or an agency manages the data collection, data management cannot continue to be an afterthought in this Age of Data.


Want to learn more about how you can make the most out of your data?
Fill out the form below: