Online Data Quality Matters
Maintaining the Highest Quality Standards in Online Market Research
How Ipsos Outsmarts Survey Fraud
In the digital age, data accuracy and authenticity is key. Ipsos’ quality systems offer a multi-layer protection, constantly monitored and updated, to guarantee the highest quality standards in online market research and detect and fight fraud in real-time. This is how we ensure that every byte of data is the unaltered voice of a genuine respondent.
Our approach focuses on three main areas.
1/ PANELISTS
We constantly verify the authenticity and credibility of our respondents and of the data they provide.
Multifactor authentication (MFA) during the registration process acts as a deterrent to fraudulent panelists. We have observed a 50% decrease in suspicious panelists and a 20% reduction in recruits using high-risk phone numbers due to the implementation of MFA.
During surveys, our real-time advanced systems detect bot-like behaviour and similarities between different accounts.
Suspicious phone numbers are automatically blocked from joining our panels.
We track the performance of individual panelists, evaluating their behaviour through a panelist score. It rewards engagement and penalizes apathy. This involves not only identifying and combating fraud but also addressing inattentive or unengaged respondents, as they can impact data quality.
Ipsos’ vetting process for external sample suppliers is rigorous. By actively benchmarking and managing partnerships, only the most reliable sources are allowed to contribute to our surveys.
2/ DEVICES
Through our advanced digital fingerprinting technology, we ensure every device has an indelible identity, at both the panel level (to avoid the creation of multiple accounts) and study level (to block people from accessing the same survey through different sources).
Suspicious devices are removed from our panels.
3/ RESPONSES
A tool Ipsos utilizes to protect against fraud is the questionnaire itself: we design certain questions to ensure we can identify respondents who may be providing false or misleading questions. We also examine the information provided in the survey responses themselves.
We automatically detect suspicious response patterns such as under- or over-clicking, inconsistencies, similarities between different respondents, speeding, and “straight-lining” (i.e. clicking on the same response every time).
Our AI algorithms detect genuine human engagement vs. AI-generated responses, especially on open-ended questions.
Despite this advanced tech arsenal, Ipsos knows the human factor is irreplaceable. A dedicated Panel Health team keeps a vigilant eye on the ongoing health of our panels, fine-tuning systems, and ensuring that only the most accurate data flows through.
In addition to the measures we take to ensure sample quality and respondent attention in our Ipsos iSay Online Access Panels, we also offer unrivalled quality with our “Knowledge Panels” in the USA and Europe, where we use random address-based sampling to ensure the highest quality in digital research combined with speed of response, as recognised by peer reviewed publications.