As ever, Cint remains focused on optimisation and innovation. After all, in the ever evolving world of digital market research, there is always much to do.
Take online fraud – a global phenomenon that extends across industries. As discussed in a recent webinar, economic pressures continue to create motivation for people to source income through any means possible; while the increase in access to reliable internet connection and the availability of devices means more and more people can join an organized group and wreak havoc in our industry. What’s more, digital market research has inherent vulnerabilities around link redirection between platforms that make it a vulnerable target.
Yet it’s important to note that the signs of fraud are different to the behaviors resulting from poor survey design and disengaged respondents. Indeed, organized fraud can be broken down into two main areas – ghost completes and script fraud.
Ghost completes occur through link manipulation when a fraudster fakes a complete. The buyer never sees the complete, but the supplier records it and pays the fraudster. In contrast, script fraud is high velocity bot automation; rapidly creating fake completes. It can normally be detected through high velocity and poor quality open-end responses.
These types of fraud are happening on a large scale, and affecting everyone in the ecosystem. Yet Cint is in a fortunate position as a global platform which sees more respondent trend data than any other platform – therefore allowing us to monitor what’s happening and to react accordingly as we continue to optimize our platform and products.
We’re also testing a new bot detection technology and leveraging the power of machine learning to proactively score the quality of a respondent. This is a learning model, remaining up to date in real time, as fraudsters adjust and change their methods of operation. I have high hopes for this as a foundational component of our platform.
We have also audited all of our third party tools. We’re optimizing where we see success and removing services that are not effectively stopping fraud. We’re constantly strategizing innovative ways to ensure safe and efficient conditions for buyers and suppliers, as we continue to deliver both speed and scale.
New methods to validate and purge poor quality panelists have been developed and these efforts will also supercharge our quality management capabilities and ensure that panels hosted on our platform offer the highest levels of integrity and quality. We look forward to keeping you informed of progress.
Yet it’s important to note that the signs of fraud are different to the behaviors resulting from poor survey design and disengaged respondents. Indeed, organized fraud can be broken down into two main areas – ghost completes and script fraud.
Ghost completes occur through link manipulation when a fraudster fakes a complete. The buyer never sees the complete, but the supplier records it and pays the fraudster. In contrast, script fraud is high velocity bot automation; rapidly creating fake completes. It can normally be detected through high velocity and poor quality open-end responses.
These types of fraud are happening on a large scale, and affecting everyone in the ecosystem. Yet Cint is in a fortunate position as a global platform which sees more respondent trend data than any other platform – therefore allowing us to monitor what’s happening and to react accordingly as we continue to optimize our platform and products.
Sophisticated automation, unprecedented scale
With Cint and Lucid having joined forces last year, we’ve also been able to benefit from even larger-scale, real time A/B testing. This enables us to identify issues and plug gaps. After all, we operate one of the world’s largest digital marketplace for research sample, with over 300 million global consumers. We’ve seen what’s been working, what’s not, and all of these insights have gone into our unified platform and upgrades, as we work to create a best-of-breed solution.Overview of quality initiatives
To give a flavour of what we’re working on, Cint is purging fraudulent panelists, validating bank accounts, checking survey open ends, upgrading third party tools, hosting panel registration checks, and securing the pay out process and survey hosting platforms: After all, it is our quest to ensure Cint delivers the highest quality sample in the marketplace.
We’re also testing a new bot detection technology and leveraging the power of machine learning to proactively score the quality of a respondent. This is a learning model, remaining up to date in real time, as fraudsters adjust and change their methods of operation. I have high hopes for this as a foundational component of our platform.
We have also audited all of our third party tools. We’re optimizing where we see success and removing services that are not effectively stopping fraud. We’re constantly strategizing innovative ways to ensure safe and efficient conditions for buyers and suppliers, as we continue to deliver both speed and scale.
Ensuring trust and driving efficiency
Our efforts have already yielded significant results. By way of example, we managed to achieve a reduction in reversals of over 50% on the Lucid platform and we’ve removed many fraudulent identifiers – over 400,000 and counting. But we’re only just getting started.
New methods to validate and purge poor quality panelists have been developed and these efforts will also supercharge our quality management capabilities and ensure that panels hosted on our platform offer the highest levels of integrity and quality. We look forward to keeping you informed of progress.
