PureSpectrum - Schedule A Demo
Our new GreenBook Directory site is live!
COVID-19 guidance, tips, analysis - access full coverage here

Are The Days Of Online Research Panels Truly Numbered?

It is essentially a Small Data versus Big Data debate. I think the multi-faceted research panels will lead to a wining Small Data strategy, likewise Big Data will play an important part in the future of research.



Editor’s Note: When we were compiling our 2016 predictions, Frank Kelly of Lighspeed GMI submitted his regarding panels. It was so good, I asked him to expand it into a whole post, and here is the result. And since next week is Samplecon, this seems like an ideal time to devote some time to disparate views on what the future of panels looks like. Frank’s post captures many of the possible paths to the future, and I think he is pretty spot on with the likely direction the industry will take. It’s good stuff.


By Frank Kelly

Research fieldwork methodologies come and go. Postal panels, central location interviewing and CATI all had their moments, but are now outmoded. By all indications, the peak of online panel research was more than five years ago when we had large, responsive, deeply profiled panels. Today, panels are less responsive; respondents do not remain in panels as long and two key benefits of panels, sample selection and panel profiling processes, have largely been replaced by lower quality dynamic pre-screening and respondent allocation algorithms.

When researchers made the transition from CATI to online research, the rationale for change was that online was faster and cheaper; however, not necessarily better. As we now transition from samples drawn from research panels to a traffic allocation method using generic respondent sourcing, the industry is making a similar decision to favor speed, cost and efficiency over research quality. Still, I think that a sustainable research panel model will yet emerge; it will just not be designed to support bulk survey completions as its primary function, but will be based upon permissioned access to a range of research information gathered from a cooperative and engaged panel. 

The Role of Panels Today

Research panels have splintered into a wide range of panel types, each dedicated to one primary purpose as described below:

  1. Online Access panels today are primarily used to get survey completions, although they are also used to find respondents for other panel types. In my classification, only panels explicitly created and marketed to respondents as research panels fit here.
  2. Purchase panels are designed to measure purchases over time from FMCG to electronics and these panels support research services for both manufacturers and retailers.
  3. Consumption and Usage panels focus on such things as the ingredients you use to prepare a meal or the face care regimen that you use.
  4. Measurement panels track television, radio and internet usage and historically have been pure currency panels that could not support other research services, but that will change, perhaps not for the currency panels themselves but other panels will offer these measurement capabilities.
  5. Community panels are similar to market research panels but tend to focus on understanding client and customer behavior. Increasingly, these will be combined with market research panels

The most obvious change facing the industry is the decline of online access panels in favor of various methods of amalgamating respondent sources and allocating them to studies, these methods now account for the majority of industry survey volume. To be clear, some clients do have a distinct preference for research panel respondents and the market will continue to provide for their needs, but a  large part of the market has moved to lower cost, lower quality methods.

Over the next five years we will see similar disruption in the other types of research panels as bar code scanning apps proliferate; we will see more competition in the purchase panel area. I believe that we will see explosive growth in various types of usage and consumption diaries as mobile devices open up new methods of data capture. Similarly, measurement panels will need to adapt to changes in the way media is consumed and distributed. As the technology behind community panels becomes more widespread and less complex, we will see a new set of competitors emerge.

Future research panels will be smaller, but will serve many purposes. What the market wants is not distinct single-purpose research panels that are cost optimized for a specific function, such as pumping out survey completes, but rather a true multi-purpose research panel that can provide a range of insights and solve a range of related business issues.

There have been many attempts at a single-source panel that captures a range of marketing stimuli and the resultant actions of those stimuli. These single-source panels were commercial successes but operational nightmares due to the cost and complexity of running them. Thanks to advances in technology and data storage, these multi-purpose panels will become more manageable and popular again. The research panels of the future will gain the access and trust of a relatively small group of respondents (in the tens of thousands) and gain access to a range of carefully curated data streams. These respondents will enjoy participating in a variety of research related tasks, from maintaining mobile diaries to involvement in discussion forums for which they are fairly compensated. They will also complete surveys, allow their media and purchases to be tracked, their debit and credit cards to be scraped along with their utility and mortgage bills.

Trust is the key ingredient driving the future of research panels; it can be earned by establishing a strong brand that treats respondents well. Trust is critical as we ask respondents for ever more invasive access to personal data.

I see the research market heading in two directions simultaneously: towards smaller, well managed research panels that attract loyal members that are willing to share their thoughts and behaviors in an environment where their input is valued and appreciated. The second direction entails large database marketing and loyalty firms that have large groups of people for which they capture data which is unrelated to research, but as a byproduct of those marketing activities, there is the opportunity to collect some useful consumer insights.

It is essentially a Small Data versus Big Data debate. I think the multi-faceted research panels will lead to a wining Small Data strategy, likewise Big Data will play an important part in the future of research, but I think those types of insights will be more closely tied to specific marketing execution. The future of panels is bright, but they need to return to what made them special; quality, representative samples, trusting and satisfied members and a diverse range of research activities. If we do not head swiftly in the Small Data direction, we risk losing the skilled people in the industry that understand how to properly build and manage research panels.

Please share...

4 responses to “Are The Days Of Online Research Panels Truly Numbered?

  1. I think the post is spot on for direction. I also believe that we have come full circle with respondents in that smaller and more focused panels can and should be composed of people with an interest in the subject matter that doesn’t require an incentive for their participation. Perhaps early access or brand advantage but not cash for time. If the engagements are succinct to the project with a defined schedule and end date, the opportunity for input stays more focused.

    Working with participants and respondents for years in the medical industry and later more consumer focused, it is apparent that people vest themselves with certain brands and certain products. They are more than willing to be a part of the brand development and their insights are generally deeper and more thoughtful. If you take a page from medical panels, where people are often searching for a better quality of life and compare it to both the depth of response and invitation acceptance rate of more general panels, you can easily see why the efficacy of general panels is questionable.

    I have no issues with large scale sampling like Google for quick questions and for general reads using census measurements. In fact, the quality is probably better because the response is not planned. Obviously it is not a good gauge for things like loyalty but it is pretty helpful in understanding if there is a need for more circumspect research.

    Panels, in their infancy were marvelous things and I worked on the forefront of that industry for many years; but the die was cast when software companies became invested. Their tech was better and less expensive but their understanding of the recruiting was poor. It didn’t take long for incentives to emerge and for the sheer volume of panels to overwhelm the individual consumer.

    Today, we have an evolved use for panels where lots of innovation, concept testing, package testing and ergonomics can be evaluated very efficiently and in some cases using extended techniques like VR and modeling for a more immersive approach to the experience. The diversity of our audiences is much more apparent and no longer ignored and the idea of pushing them into pseudo-stat buckets is decidedly less appealing and without question a less accurate representation of the consumer population.

    The advent of social media has also made audiences more attainable and the ability to invite participants from forums creates a higher confidence in the investment of the respondent in the subject matter.

    There is really no need to maintain large panels and deal with attrition, incentives, etc. when there is access to qualified individuals. Even the use of customer panels is changing from a research application to a more interactive model where cohorts are “activated” from a database for their specific purchases or experiences. It’s pretty obvious these days in retail when a store is not performing, the metrics are real time. We aren’t doing discovery anymore, we are honing in on specific issues.

    The elephant in the room for research is that consumers see brand interaction as a two way street. They are often ahead of any research in their posts and their ability to create con consensus on an issue or virility on a find. They aren’t politely sitting around waiting to be asked. The other obvious point is that people are pretty good at judging reasonable response. The fluidity of online focus groups, forums and short term panels provides a pretty good way to stay “in the moment” with customers.

  2. Great article. Importantly we need to think about the members of the panels, as an industry we must continue to think about the experience of receiving a survey, completing a survey and what we are really trying to understand from the member (collect key insights, not just a lot of data that will never be used!). We must move to shorter, interesting, mobile friendly surveys. If you or your mum would not complete the survey, why on earth do you think a panel member would complete the same survey?? And, BTW flash in dead!!

Join the conversation