Skip to content

Blog

Data Quality: Adding the Next Line of Defense

Data Quality: Adding the Next Line of Defense


As a research logistics company, quantitative data collection is a huge part of what we do. We have always had a strong commitment to data quality and eliminating fraud, but as fraudsters get more innovative, we have to adapt as well. We want to take you through some of the ways that Ag Access has always protected your data, as well as some new tools we are utilizing to take it one step further. 

What We've Always Done

We have always had a thorough set of data quality checks that allow us to remove participants that appear fraudulent or to just not be taking our research seriously. These checks include looking for: 

  • ReCAPTCHA verification
  • Bad IP/Domains, including ones we’ve previously determined to be high risk 
  • Fake personal information 
  • VoIP phone numbers 
  • Contradicting answers or answers that don't make sense - for example, growing a crop where that crop cannot be grown
  • Gibberish or off topic open end responses
  • Straightlining - answering the same response for every question on a scale
  • Speeding - moving through the survey too quickly
  • Operation location and pay to locations that don’t match - someone says they farm in California, but they want their check sent to Florida
  • One person/IP filling out the same survey under multiple different names


Whenever a respondent is flagged for one of these, our team conducts a thorough investigation to determine if the respondent is truly fraudulent and warrants removal. 

What We're Adding

While we have always been committed to data quality and reducing fraud, there is always more we can be doing, and that’s why we’re making some changes. We decided to move forward with implementing a data quality software called Imperium as well as pursue ISO 27000 Certification

Imperium

What is Imperium? A software designed to verify personal information and reduce the probability of fraudulent activity within market research data collection. It uses machine learning, natural language processing and neural networks, as well as various other factors to identify fraudulent patterns. 

Imperium has many parts, and we have begun with the implementation of RelevantID which verifies respondent identity as well as RealAnswer which validates open ends. In the future, we look to expand our usage of Imperium with additional solutions. 

RelevantID

RelevantID gives each respondent a fraud probability score based on various factors. We terminate automatically based on fraud probability score and/or duplicate score of 90% or higher, and manually evaluate each record with a 70% or higher score at either variable for fraud risk.

Respondents are flagged for reasons such as:

  • Spoofing - mimicking the properties of another device
  • Suspicious IP addresses or domains - flagged for high fraudulent activity in the past
  • Bots/crawlers - programs or machines designed to find and enter survey links
  • Country/location mismatch - device identifies as one location but IP pings in another
  • Duplicate entries - based on device IDs, cookies, verifying the “unique fingerprint” from each device


RealAnswer

RealAnswer checks open ends for things like bad/garbage words, unrelated answers, robot submitted responses, copy and pasted responses, repeated words, and duplicated responses. While these are already things that we do, having RealAnswer added to our surveys allows this process to be automated and free up team member time to focus on other efforts that require more of a human touch to move projects forward. 

ISO 27000 Certification

Ag Access has strong business practices, personal accountability, and technical security. We are pursuing an ISO 27000 certification, which is a global certification related to information security when collecting personal information, to formally validate our commitment. Every member of the Ag Access team is participating in extensive training over the course of two years and the company will participate in an external review and audit of our systems prior to final approval and registration. Following the completion of the certification, Ag Access will continue to be audited each year to maintain compliance and prove our continued diligence. 

What This Means for Our Clients

Our efforts are already helping. Community members go through a profiler survey to join, and our data quality practices are helping weed out fraudsters before they even get into the community, meaning that they never even touch live surveys. Since implementing Imperium, the Ag Access community has had 1020 registrants and Imperium has stopped 4% by marking them as fraudulent. Our sister brand, Vet Access, has had 917 registrants, and Imperium stopped 10% for fraud. Because we stop these people from getting into our community in the first place, we have a lower rate of fraud on individual surveys - only about 1% of respondents on average.

We are a small business built on the relationships we have with our clients and customers, but by utilizing our internal resources and this large software platform, we can still protect data like the big companies. When we discussed our new system with one client, they said "That sounds like a pretty sophisticated system - I like it. The scammers are getting too smart and are using technology, and we’re having to scramble to keep up with them. So, I appreciate a good system like this!"

The Impact of Talking to the Right People

When you know you're talking only to real people with good intentions, you can be sure you're getting real results. But talking to the right people get you the right results to enable you to make the right decisions. Check out our case study Access to the Right People Can Shift Your Decisions for Good to learn just how impactful this can be.