Overcoming the challenges of data quality in panel research

aerial view photography of group of people walking on gray and white pedestrian lane
aerial view photography of group of people walking on gray and white pedestrian lane
Photo by Ryoji Iwata on Unsplash

Market research has undergone significant changes in the last few decades. It wasn’t long ago that researchers needed to recruit in-person focus groups, or conduct nation-wide paper surveys to find the answers they needed to better understand their customers.

The digital age created a new and exponentially more efficient path to consumer research in the form of online survey panel providers. With so many players now in this space, it’s important to remember that not all panel companies are dedicated to delivering research outcomes at the highest level. It is not enough, and not acceptable to focus solely on cheap pricing or panel size at the cost of panel quality.

The danger of this kind of panel approach is that it can lead to organisations basing their decisions on irrelevant or incorrect data and this can have significant implications down the line. For example, if the panel provider doesn’t screen for authenticity threats such as bots, fake accounts or false responses, the results produced will not be an accurate representation of consumer sentiment.

Robust, tangible research can only be based on the thoughts, behaviours and opinions of real people. This is why it’s critical to always work with a research partner who manages their own proprietary panel community – cheaper alternatives such as river sample or panel aggregator services are more likely to lead to misinformation.

Pureprofile has been at the forefront of the digital research space for over 20 years – this tenure in the industry has taught us some very valuable lessons about procuring data quality.

While we’ve fine tuned our systems, measures and methodologies to provide the highest quality research outcomes possible, we’ve identified three critical tips for you to use before starting your next research project:

1. Always investigate the authenticity of your sample

It’s crucial to ensure that the panel you’re using is authentic and reliable. Your panel provider should have a well documented member validation procedure in place. Panelists must go through a rigorous process at the registration stage, including email address, mobile number and identity validation. In addition to this, regularly tracking device and location data while looking for inconsistent survey responses can help to assign a scoring system for each member – helping to identify situations that need closer monitoring.

Some of the validation techniques that Pureprofile employs includes IP address matching, name matching, password matching, and unique machine ID validation. It’s important that there’s a stringent policy in place, so that respondents who are identified are given chances to provide quality data. If they fail to do so, they should be removed from the panel entirely.

It’s important that panelists are aware of what is required of them to be a part of a research community. Due to the monetary nature of a survey reward system, clear ground rules need to be established and communicated so that panelists understand that any shortcuts or attempts to trick the system will eventually result in disqualification. For example, it is expected that new Pureprofile panel members complete our “Education series” – an interactive, step-by-step guide to panelist dos and don’ts. Not only does this provide an engaging way to let our members know that their honesty and accuracy is mandatory, it gives our panel care team a clear indication of when each panelist agreed to these rules, to be referenced at a later time if necessary.

2. Remove all bias from your survey design

We all have inherent and unconscious biases, so it’s critical to ensure that this doesn’t appear in your questionnaire. Was the survey design significantly different from earlier research that was done previously? Has market share or brand awareness been defined differently? Should you word a priority question as an open-ended question or close-ended question? Surveys need to be consistent and flow seamlessly for respondents. It’s important to ensure that each question is clearly written – removing any leading questions or biased answer sets.

The best way to remove bias is to conduct a proper questionnaire review. By asking a handful of trusted colleagues or acquaintances to be a “respondent”, you can watch them go through each step of the survey and determine any gaps. Were there any points that they couldn’t understand? Were there too many repetitions of similar themes? Do any of the questions sway the respondent to answer one way or another? By implementing these checks and balances, you can give your research the best possible chance of being an accurate representation of consumer sentiment.

3. Create a positive and engaging survey experience

Consumers are bombarded with requests for feedback and survey invitations – even after showing brand loyalty and purchasing a product online. Respondent fatigue is a genuine issue that needs to be addressed. One way to captivate an audience is to use attention grabbing or knowledge-based survey questions to ensure that respondents are engaged. Additionally, sharing encouraging messages at multiple stages within the survey will ensure that respondents know they’re getting closer to completion.

Your panel provider should provide you with a detailed account of how they approach panel care. At Pureprofile, member experience is at the forefront of everything we do. We know that if our panel members are presented with interesting surveys that are relevant to them, the more likely they are to provide authentic responses.

Our dedicated Quality Assurance team exists for this very reason. This taskforce is made up of experts that bring a wide range of experience to the table including; established programmers and scripters, project managers who’ve led studies across every possible vertical, in-house data scientists, panelist experience managers, as well as account managers who are dedicated to representing the exact outcomes a client requires. This combination of technical and industry know-how ensures that our processes are constantly optimised to meet and surpass the ever-evolving demands of the research sphere.

By paying particular attention to not only the robustness of the sample itself – but also how balanced and engaging the surveys are, brands, agencies and organisations across the board will be able to avoid the most common factors that lead to poor quality research.

Ultimately, the purpose of market research is to learn what really matters to consumers. This requires careful consideration, the right panel partner, and the understanding that impactful research can only be achieved with a rigorous approach to data quality.

By Keith Ang, Client Solutions Director, Asia at Pureprofile

This article was first published by Research News Live