A few weekends ago I shipped a new feature that helps you better understand and serve your followers on Twitter. It’s called Follower Intelligence.
Follower Intelligence tells you:
- WHAT your followers are talking about
- WHEN your followers are talking
- HOW often your followers discuss the things they care about
You can combine this intelligence with the tweet scheduler to optimise marketing activities. For example – by scheduling contextual tweets that correlate with peak times in the heatmap you can ensure the right message is received by your followers.
Under the Hood
Machine learning models identify the main places, people, products, or services that your followers are discussing. For example, I recently shared a tweet asking for opinions on a new Android phone (mine is 4 years old!).
I have a development account on Twitter which my personal account (above) follows. Here we can see tweets from followers of my development account. The tweet from above is included.
We can see in the above screenshot that further products, brands or services have been identified in tweets being sent by followers of my development account.
Visualising This Data
A lot of accounts on Twitter have thousands of followers and you need a quick way to find signal in the noise. This is where the tree map and heatmap help you.
Tree Map
The tree map shows you WHAT your followers are discussing. Data is grouped into give categories which are:
- Products
- Organisations
- People
- Places
- Miscellaneous
Within each category the top instances of each type are displayed. In this example, we can see what my followers have been talking about in real-time:
Note: Places/Misc are not included as there is no mention of those from my followers at the time of writing
You can hover over each segment in the tree map to see how many times a particular item has been identified by the machine learning model. Being able to quickly see what your followers care about is good to know.
Heat Map
The heatmap shows you the WHEN these tweets were created discussing the above data. The volume of tweets is also displayed in each cell. We can see that my followers tend to be more active in the red zone (Wednesday to Saturday and mainly between 1400 and 2300):
This makes sense as most of my “audience” whether it be on this blog or from my Twitter account tend to be from the US. People in other time zones may also have finished up work later in the evening.
Use Cases
There are many use cases for these insights. Here are some examples:
Content Marketing – create content focused on topics your followers find popular.
Audience Intelligence – know when your followers are most active & schedule content at these times.
Boost Engagement – increase your reach and impact by knowing the best times to engage.
Trends and Patterns – report upto 120 days in the past to identify trends or patterns.
Summary
In this blog post we’ve introduced the audience intelligence feature. We’ve seen the data it provides, explored the visualisations, and touched on some of the use cases.
You can try this out by visiting www.socialopinion.co.uk
Leave a Reply