It’s true that not everyone uses every social media platform, so we commonly get asked how we handle bias.
Top myths about social media bias
- Social media users are young: Not any more. Over 60s is the fastest growing group of social media users, and the average Facebook user is over 40. We have tested our data extensively in communities of older Australians, and there is no statistical drop off (in fact, it's very common to see the local RSL, or Bowls Club coming through as the most loved place - point proven!)
- Social media is only used for recreation: If you think that social media is only used for sharing cat videos, or posting holiday snaps, then you're wrong! In Australia, more than 95% of businesses with an internet connection maintain an social media presence of some kind.
- Lots of people don't use the internet: Covid really changed this. The move to remote service provision and preference for digital transactions rapidly digitised our local economies. These days Australians spend an average of 6 hours and 14 minutes on the internet EVERY DAY.
- Whole segments of people - such as homeless communities, or CALD communities - do not use social media: Very untrue! Social media is an increasing channel of choice for many marginalised communities that may have once been seen as on the periphery of conventional communications. The peer-to-peer nature of these platforms commonly makes them more quickly trusted than traditional communication channels
How Neighbourlytics handles bias
It’s true that there are individuals in any demographic who are not online, but what’s not true is that there is one total vertical missing.
In fact, when looking into the demographic bias that comes with social media, the biggest skew is not 'who is online', but rather 'how do people engage with the platforms?'.
To generalise, younger people are found to check these platforms many times per day, moving across large quantities of content, but not interacting. While older people check less often, but for longer sessions and are more inclined to interact and leave comments.
For this reason, there are two key ways that Neighbourlytics reduce the bias of our data;
FIRST. We tap into multiple different types of data, from a number of different platforms, to ensure that we’re covering all kinds of different lifestyle behaviours.
For example if we only used reviews, it would skew 'older', or if we only used impressions, it would skew 'younger'. But by bringing in likes, checkins, reviews, posts, impressions etc... essentially all the digital activity that we can access, we are able to appropriately weight the results to reduce bias.
SECOND. We compute the whole country, every day. It is much easier to quality check and spot gaps or bias behaviours across large geographic areas, than if we were undertaking adhoc analysis.
This means we are regularly checking and adjusting our computations to make sure that we are identifying and handling biases appropriately.