Vox Pop: Is social media intelligence and digital consumer intelligence eroding traditional respondent sampling to "respondents of convenience"?
When it comes to social data there have always been discussions about whether the sample is representative of a population, and if that really matters. This week, we asked our experts if social media intelligence and digital consumer intelligence was eroding traditional respondent sampling to "respondents of convenience", and what this means for research bias and representativeness? Here's what they had to say...
Tara Beard-Knowland, Head of Social Intelligence Analytics, Ipsos
We hear sometimes that ‘it doesn’t matter, everyone is online’ in research. That is frankly untrue. The best in class political polls are still face-to-face exit polls. None of the big media measurement contracts (like BARB and RAJAR in the UK) are done fully online or entirely with a DCI audience (they’re largely face-to-face still). Online audiences (even on panels) are not truly representative. When you want to know a representative view from the public on an issue, DCI will not cut it as the only method.
That being said, social intelligence (or DCI) still has a great deal of value, even if it will not give a fully representative view. In the first instance, when it comes to public debates, social intelligence is brilliant at helping us understand polarised viewpoints. Also, in standalone research, there are many times when a representative sample doesn’t actually matter – whether you’re trying to discover new product ideas or to understand purchase journeys online or any of a myriad of other topics.
DCI is not eroding typical respondent sampling to ‘samples of convenience’, but at lower prices and easier access, DCI is making some organisations consider just when they need the fully representative data sets. The vast amount of data and the agility that digital methods give us have also made the 70% or 80% answer be just what organisations need. So, yes, DCI has changed the way people look at information, but that is not necessarily for the worse.
Tim Wilson, CEO, Qutee
This evokes a broader question of what is the perfect sample? The perfect piece of research? Social listening can provide a widescreen view but lacks the granularity and accuracy a focus group can provide while focus groups lack the scale and speed. Meanwhile, how can panels ever compare in terms of passion and knowledge to an influencer's audience or brands own consumers?
For example, if you're looking for insights within the auto sector, then undoubtedly utilizing car influencers to drive the research has to be a consideration?
What if you can create your own industrial scale focus group in the 10,000's via your CRM or social channels and drive strategically relevant research daily? Surely those insights would have more strategic relevance, be the least bias and have more value than the existing funnels of insight?
Want to drive more comprehensive research we'll use a broad spectrum of influencers who have expertise in the sector.
In my opinion, we are entering the jet age of consumer research where any SME's social manager will in hours be able to drive consumable and actionable insights on a par with the major research agencies for a fraction of the cost.
Javier Buron, CEO, Audiense
From our experience, social media intelligence should help traditional consumer research to test hypothesis and create a more effective traditional market research. In other words, social media intelligence should be taken as qualitative research and used in combination with traditional respondent sampling.
Another important strength of social media intelligence is:
- The ability to benchmark (this surfaces the real insights by removing a big part of the bias).
- Segmentation by interconnectivity.
- The understanding of the affinities towards brands and/or influencer of people without bias.
These three elements that more traditional respondent sampling doesn't have makes social media intelligence provide unique answers.