What social media analytics can tell us about elections
What did social media last year predict about the 2016 federal election? It depends on how you look at the results.
What did social media analytics last year predict about the 2016 federal election? It depends on how you read the results.
Quiip monitored mentions of Tony Abbott, Bill Shorten and Malcolm Turnbull and associated social media accounts for four weeks from Sunday, 18 October to Saturday, 14 November 2015. On 15 November 2015, we reviewed the social media analytics for each Australian politician.
In a surprise to many, Abbott was the most popular politician on social media. Just look at the results:
Social media fans:
Social media mentions:
Abbott was the most popular of the three politicians monitored. He was mentioned 1.2 more times than Turnbull and an astonishing 4.6 times more often than Shorten.
You may have noticed I’m being a bit tongue-in-cheek. Data is one component of a beneficial social media report, or any report for that matter. Data integrity, data interpretation and tool selection are important for creating valuable social media reports. You need to select the right social media tools to meet your needs. You also need to make sure your staff knows how to use those tools and get the most out of them.
Turnbill had the most positive sentiment and the highest reach.
So, was Turnbull the most popular of the three?
Sentiment has been automatically calculated and can only be used as an indicative measure. It needs more context to be reliable. This report contains a small sample of metrics that can be used to report on social media activity. It’s important to interpret data using varied metrics to get a contextual and reliable report on any topic.
If we looked at different metrics or read them differently, we could also argue that Shorten was the most popular of the three politicians.
Many organisations see sentiment as a key measure of social media efforts, but sentiment reporting is not an exact science. To improve the integrity of this data, compliment tools with manual sentiment classification.
What do positive, negative or neutral sentiment really show?
It’s important to compliment data with examples of each sentiment classification. Even then, what do the different classifications mean? Positive can account for happiness, optimism, excitement or any number of emotions. Negative sentiment can cover anger, sadness or disappointment. Positive and negative don’t do justice to the spectrum of reactions and emotions expressed hourly on social media.
We read data for four weeks from 18 October to 14 November. Early during this period, Abbott made a speech in Europe about refugees. A barrage of critical news headlines and social media sharing followed. Mamamia said on Twitter: “Tony Abbott says compassion towards refugees is a catastrophic error.” How would a sentiment tool classify that tweet?
Context is important. For example, mentions of Turnbull exceeded mentions of Abbott during 8-14 November. If we’d selected that data, we could have created a report favouring Turnbull.
Context is also crucial when building social media reports. Abbott was Prime Minister for two years. He had a higher profile and more time to build a social media following. Abbott was controversial PM. His speeches towards the end of his Prime Ministership were closely followed and reported on.
Turnbull became PM in mid-September, a few weeks before us we started tracking searches. He had a shorter period to build a social media following than Abbott. And what of Shorten? The leader of the opposition had a much lower profile than Abbott or Turnbull. He received less news coverage and social media commentary.
Great social media reports are an art we’re still refining. Worthwhile reports discern not only data. Good reports:
Are you getting the most of out of your reporting? Are you tracking the right analytics? Are you interpreting the data properly? Discover a new approach to social media monitoring, analytics, and reporting. Click here to find out more about Quiillion.
We'd love to hear what you have to say in the comments belowComment & Discuss