The Global Source for Social Media Researchers

The Future of Research is Social

18 Feb 2021 9:32 AM | Kathy Doering (Administrator)

Published in Quirk's Magazine Jan.Feb/2021

Rodolphe Barrere, Co-Founder and CEO, Potloc 

With 3.96 billion people using social media today, sampling on social channels is the next research frontier

According to Hootsuite’s latest Digital Statshot Report, more than half of the world’s population is now on social media. Not ever in human history have we had such vast access to information, people, news and interconnectivity. The implications of this are numerous in terms of human communication, how people interact with brands and businesses worldwide and how companies extract consumers’ insights. 

When I started building Potloc in 2014, I recognized the power social media had to gather people’s thoughts about what was going on in their immediate vicinity. We helped local businesses identify what offering, location and combination of customer experiences would appeal to their trade area. It didn’t take us long to realize we were sitting on a real gold mine of insights. The samples we were getting through social channels were on point, with relevant respondents answering our surveys with no other incentive than their desire to be part of the conversation. Through testing and trying different approaches, we soon discovered that we could leverage the targeting capabilities social platforms offer to reach any person with a social account, anywhere in the world. This realization allowed us to scale our operation and offer our services anywhere you could find someone “thumbing-up” through their social feed. This approach might seem obvious but the reality is that sampling in social media as a methodology has been largely overlooked as a great way to reach niche audiences and fresh respondents on a global scale.

The medium is the message

The key to nailing social sampling boils down to really understanding the medium. Social channels’ dynamics are quite different from those of very controlled environments like focus groups, online panels or even face-to-face interviews. To gain insights from social users, you need to understand how the social ecosystem works: people use social media mostly when they have downtime – moments of free time they dedicate to browsing their feeds, a golden opportunity to interact with them. On average, the world’s internet users spend two hours and 24 minutes using social networks across all devices each day, accounting for more than one-third of our total internet time. “Intercepting” them during this time increases their chances of taking a survey and focusing on expressing their opinions. Non-intrusive research and how companies approach potential respondents also plays an essential role in conducting successful sampling. The messaging not only needs to be appealing to them but also relevant and at the right time. After more than seven years of developing our expertise in social media sampling and the technology that enables it, we know how to increase the relevancy factor. Here are my thoughts:

Location, location, location

Social media is the perfect place to reach respondents, no matter the incidence rate or the quotas you have set. The beauty of targeting people via social networks is that you can reach them at a broad location level (country, state, zip code), down to a half-mile radius. Through geotargeting or more conventional location-based targeting, businesses can reach highly relevant audiences. People who live, work, study or just transit through the targeted area are real-life respondents that have a treasure trove of insights to share, from in-store CX to people’s experience attending an event or having seen a particular billboard at a specific location. The possibilities are endless.

The right targeting can get you niche audiences with low incidence rates

Again, the nature of social networks is that people are there to share. Think about it – when you browse through your feed on any of these networks, you are already interacting with brands and businesses to a certain degree. This gives researchers unprecedented opportunities to access the largest consumer group globally and extract insights directly from them. The essential advantage here is the ability to target a specific population, not only by geographical area, as I mentioned before, but also by age, gender, interests, language, profession, income, etc. We’ve worked for the largest consulting firms globally, like Bain & Company, BCG and EY, to help them reach very niche and specific audiences for their clients, where panels tend to struggle. Hitting the bull’s-eye with targeting on social guarantees a more representative, diverse and genuine sample of the broader population, with fresh respondents that reflect real people living in the real world.

The biases to keep in mind

Just like any other methodology, we deal with some biases that come with social sampling. We have identified four of these as well as ways to tackle them:

  1. Coverage bias: For targeting consumers on social networks, they need to meet certain conditions. They must have access to the internet, have a social media account and be active users. This might show an under-representation of men and older people. However, older generations like Baby Boomers and the Silent Generation have adopted social platforms in great numbers. In the U.S. alone, 72% of Americans between the ages of 50 and 64 are on Facebook and 62% of “online seniors” aged above 65 are also there.
  2. Ad platforms’ algorithm bias: Taking Facebook as an example, its advertising algorithm is set up to minimize cost-per-click. This means it pushes survey ads primarily to the least-expensive audiences (under-representing the population). Setting up quotas to balance the sample becomes essential to counter this bias, especially when it comes to more expensive-to-reach people, like men and the elderly.
  3. Cognitive load bias: Answering an online survey is demanding from a cognitive standpoint, so some people might find the task too difficult to complete, which gets exacerbated by the use of mobile devices. This might result in an under-representation of older people, less-educated or illiterate people or people of a low socioeconomic status.
  4. Self-selection bias: Unlike web panels, with social sampling we have to communicate on the survey subject. People who click on our ads are more inclined to provide answers about a specific topic. As we rarely offer any incentives to respondents, people who complete our surveys do it because it matters to them that their voice is heard. In my opinion, today it’s hard to say what is the lesser evil: having respondents naturally interested in the subject or respondents seeking incentives.

Reaching the unreachable

As we witness a new world order emerge as a consequence of the COVID-19 pandemic, surveying people now comes with a set of limitations. Social distance and lockdowns are a no-go scenario for in-person or intercept interviews. People are spending more time than ever online and at home and a third of that time on social media. Again, social sampling here presents itself as a unique opportunity to reach people in impossible places and situations under these circumstances. At Potloc, we launched twin studies in Canada and France about what frontline health workers saw in the trenches against the virus. We reached them in emergency rooms, hospitals, nursing homes and places no one else could enter, at a time where it seemed impossible to get their insights – and with great success. Niche and low-incidence populations are out there, browsing their social media feeds. We just need to find them and offer them the chance to express their opinions.

Interested in learning more about social sampling and how we run respondent acquisition at Potloc? Check out our public studies at or better yet, e-mail us today to schedule a discussion with one of our research experts:

  • Home
  • News
  • The Future of Research is Social

Copyright 2017, Social Media Research Association. All rights reserved

Powered by Wild Apricot Membership Software