Social media population bias distorts big data
Marketers planning to use social media as a means of mining data about their target audiences have been warned about population bias.
Social media is seen as a hugely valuable resource for marketing teams, as it provides low cost access to the thoughts and opinions of potentially millions of people.
But a recent study published in the journal, Science, has cautioned that leading social media sites often have a significant bias towards certain demographics, which can distort the results, particularly when it comes to large datasets.
Researchers used Pinterest as an example. The visual sharing tool has more than 70 million users globally, but it is disproportionately popular with women between 25 and 34 with average household incomes of $100,000. Data from Pinterest would therefore suffer from sampling bias, no matter how big the sample.
The researchers also noted that some social media profiles are inactive or fake, while others are run on behalf of brands rather than individuals. In some cases, the same person might have multiple profiles on the same social site. These factors will also have an influence on the reliability of social media data.
But while big data may be undermined by sampling bias, social media can still be a useful resource if your ambitions are a little more modest.
For example, if you’re trying to identify topics for your blog, listening to conversations on relevant social media sites can often be a good way to tap into what your target market is interested in.
If you want more 24 to 34-year-old women with above average income visiting your website, you could certainly try blogging about popular topics on Pinterest.
This kind of flexibility when it comes to your content marketing strategy can be really useful. If you can stay nimble and react quickly, you’re much more likely to be able to give your audience what they want, which will earn you convertible traffic and more of it.