Effective
social media optimization is great for building brand awareness and getting people excited about your offering.
To be effective, your social media posts need to spark a lot of engagement. Which raises the question, why do some posts drive more engagement?
It’s not an easy question to answer, but if you have the right social media optimization tools, you can see what works and why. Then apply the learnings to achieve higher engagement with new posts.
How do you know which posts will drive engagement?
The tools behind NEURO FLASH’s
social media optimization service are, you guessed it:
- Artificial intelligence: to predict how content is being perceived by a typical human being
- Machine learning: to predict which associations lead to more engagement
- Data visualization: to tell the story in a way that humans understand
Why? Because these tools are capable of handling the big data that comes along with using such large social platforms. Then, using the information you analyzed, you can fine-tune your social media optimization (SMO) strategy to include more of the things that work while avoiding the things that do not work.
To illustrate, you will read in this post how we used these tools to analyze UNICEF’s social media posts. Plus, we will highlight the major insights we discovered, which can also help you create more engaging content.
Note: We did this analysis without UNICEF directing our efforts. UNICEF did not pay for this analysis nor did they influence the conclusions of this post.
Before we dive into the process/methodology, you need to understand associations. Why? Because associations (the automatic things people think when they see something) are driving the impact of your content.
What are associations?
Short Background on Associations
Associations are the automatic connections we make between words, concepts, and ideas. They are a basic phenomenon of how the brain works. There are two types of associations:
- Explicit Associations: The associations you can voice and point out when prompted to do so. For example, if I ask you ‘How does the color yellow make you feel?’ your response reveals your explicit associations as you are immediately conscious of them.
- Implicit Associations: The connections we make internally without outwardly realizing it. Thus, implicit associations occur below your immediate consciousness and are predominantly subconscious. So, if I ask you ‘How does the color yellow make you feel?’ and you say ‘Good.’ But, you pause/hesitate before answering without knowing why, that’s an example of an implicit association. As a rule of thumb: “When you aren’t aware, it’s implicit.”
Example of Associations Using Data Visualization
Here is another real-world example of associations, but this time illustrated through data visualization 🙂
Suppose a man and a lady are having a date at a restaurant and the waiter brings over their drinks. Without asking, the waiter places the beer in front of the man and the lemonade in front of the woman. Has this ever happened to you?
Why did the waiter do that? Because his brain naturally and more easily associates man and beer.
This is the power of associations, they influence not only the way we think but also the way we act.
Interestingly, consumer associations can be captured using big data by collecting all of the billions of words openly available from culture. Then, these data points can be combined with machine learning to predict the associations that consumers make. This is what we are all about.
Using data visualization, we demonstrate with the green connections between man and beer, that there are a lot more associations between man and beer than between man and lemonade – which can make it a lot easier for most people to automatically connect/give beer to a man rather than a woman.
Associations in the Digital World
In the digital world, you can analyze all of your social media posts and identify which posts were more or less successful.
Then, you can deduce which associations drive success, allowing you to be aware of which associations to use in new posts. You can also see which post associations were poorly linked to success, thus enabling you to avoid them in the future.
This is the essence of Social Media Optimization (SMO). By using social media optimization tools like big data and machine learning, you will get insights that can effectively guide your future content creation. You will see what we mean if you take a look at the rest of this article 🙂
Now that we’ve briefly reviewed what associations mean, let’s see how we used them to understand what drives the success of UNICEF’s social media posts.
UNICEF’s Social Media Optimization: The Process
We will take a look at UNICEF’s:
- Social media output
- Analyze which posts are more successful
- Determine WHY those posts are successful
Social Media Optimization for UNICEF’s Facebook Channel
The topic that UNICEF is tackling isn’t easy to write social media posts about. It’s quite a challenge actually. How do you boost engagement when your topics are dark?
Famine, disease, migration, and malnutrition aren’t your typical clickbait topics. While the challenge is daunting, the importance of raising awareness is equally big.
So, we accepted the challenge. We set out to determine how UNICEF can increase their social media engagement while raising awareness of important global humanitarian topics.
Which associations are connected to more engagement for UNICEF?
Overall, the questions we had to answer were: Is it better to have people think positive thoughts or negative thoughts. Which overarching theme is more powerful in driving social media success, hope or despair? Which side leads to more engagement?
Data Collection
We collected data from UNICEF’s Facebook page:
Overview of data
For later steps in the analysis, we will look at post success. And we have defined success as a high engagement score as calculated the total number of likes, comments, and shares, i.e. if a post had more likes, shares, and comments, it got a higher success score.
Base Analysis: Methods
For each Facebook post, an analysis was performed using an embedding model. By doing so, the top 10 words for each post were identified. Then, all the top words across all 2971 posts were collected.
Aggregated top words identified by using embedding models on UNICEF’s 2971 Facebook posts
After identifying the top words of each post, they were clustered, based on their semantic similarity. i.e. posts that were more similar, were clustered together.
Just like the man-beer example, when words connect with each other, they are more closely related in people’s minds. With this approach, we determined topic clusters.
Three major themes were:
Visualization of how words are used to form the 3 semantic clusters. Note how words at the border between clusters are also semantically “in-between”.
From all of the 2971 UNICEF posts analyzed, there are 3 major clusters/themes: Family, Refugees + Migrants, and School + Education.
Top 3 clusters identified by our embedding models based on top words
Text Analysis: Which associations are triggered, that lead to higher success?
Before exploring the 3 clusters individually, all of the posts were analyzed together. Thus, we were able to identify the general factors that contributed to the success of social media posts for UNICEF.
Correlations of associations invoked in successful posts (represented in yellow bars) and less successful posts (gray bars)
Our goal here was to show which associations were predominant in the most successful posts (represented by the yellow bars). That means that the following thoughts were created by the post content. Thus posts that elicited these associations were more successful/engaging than average posts.
Text Associations with Higher Engagement
The following text associations were found in the most successful posts:
- Glowing: shine, glow, glint, glisten, shimmer, bright, moonlight
- Musical: sing, rendition, music, vocal, croon, lyric, chorus, singer
- Cheerful: wonderful, joyful, wonderfully, fabulous, exuberant, lovely, delight, amaze, cheerful
- Goofy: cartoon, kooky, lovable, funny, clown, adorable
- Grateful: thank, thankful, grateful, wonderful, 🙂, awesome, happy, congrats, appreciate
By knowing which associations are linked to high engagement, UNICEF can set out writing future posts that are predetermined to align with the successful associations.
Text Associations with Lower Engagement
And, in the least successful posts, the following associations were invoked (represented by the gray bars):
- Illness: illness, symptom, disease, acute, sufferer, pneumonia, debilitate, ailment, syndrome
- Vaccination: influenza, diphtheria, malaria, vaccination, measle, vaccinate, hepatitis_b, dengue, h1n1
- Childbearing: infant, maternal, neonate, childbearing, childbirth, pregnant, neonatal, adolescent, fetus
- Malnutrition: famine, starve, poverty, malnourished, epidemic, hunger, malnourishment, impoverish
- Medical staff: medical, hospital, patient, doctor, pediatric, nurse, medicine, clinician, pediatric
What does this mean for Social Media Optimization (SMO)? It means that now you know which associations are linked with engagement. So, for future posts, we recommend invoking these associations in order to boost engagement. Thus, big data and machine learning (which make this analysis possible) are insightful social media optimization tools that can inform your decision-making for better content creation.
Examples of Posts with Higher Success Associations
Below, you can see how the highly successful associations (like Cheerful, Glowing, and Musical) are identified and measured in real-life Facebook posts.
The yellow highlights are used to visualize the strength of the goal association and the word. This is made possible via the in-browser
NEURO FLASH chrome extension. The extension works while you are logged in to your account. For your free 3-day trial,
sign-up here.
‘Cheerful’ Association
‘Glowing’ Association
‘Musical’ association
With such insights, you can craft your posts with real-time feedback that informs you whether your text is aligned with your target associations or not.
Examples of Posts with Lower Success Associations
On the other hand, here you see in grey the associations connected to low success:
‘Malnutrition’ association
‘Childbearing’ association
‘Medical staff’ association
These posts and their respective associations scored poorly and had a low engagement. These are examples of what to avoid. With advanced social media optimization tools such as this, you can fine-tune each post so that it definitely sends the message that you want to send.
How to apply these learnings for YOUR future social media posts
Equipped with the associations behind more successful content, you can now create new content that effectively triggers those “winning” associations. To that end, we have created the TESTER. This online software allows you to predict the associations your content triggers and helps you optimize marketing content, like slogans, headlines, and even social media posts.
Take a look at our
free-trial offer. Or take a look at another use case, how to enrich your headlines with brand values,
which you can find here.
Back to the UNICEF analysis.
Image Analysis: Which Image associations are triggered, that lead to higher engagement?
Correlations of associations invoked in successful posts (represented in yellow bars) and less successful posts (gray bars)
Image Associations with High Engagement
The most successful images invoked these associations:
- Cinematic: film, filmmaker, actress, screenwriter, documentary, oscar, filmmakers, biopic, motion_picture
- Legendary: famous, pay_homage, legendary, heritage, masterpiece, timeless, contemporary, renowned, tradition
- Treaty: agreement, negotiation, pact, ratify, signatory, ceasefire, ratification, declaration, multilateral
- Adventure: embark, journey, adventure, treacherous, adventurer, adventurous, explore, unexplore, expedition
- Natural disaster: torrential_rain, tsunami, typhoon, hurricane, flood, earthquake, torrential, rainstorm, cyclone
Image Associations with Low Engagement
As shown above, the image analysis revealed that, overall, the least successful images had the following associations:
- Trauma: emotional, physical, mental, trauma, anxiety, depression, cognitive, physiological, emotion
- Malnutrition: famine, starve, poverty, malnourished, epidemic, hunger, malnourishment, impoverish
- Vaccination: influenza, diphtheria, malaria, vaccination, measle, vaccinate, hepatitis_b, dengue, h1n1
- Unfortunate: terrible, frustrate, wrong, unfortunate, sadly, horrible, awful, badly, spite
- Medical staff: medical, hospital, patient, doctor, pediatric, nurse, medicine, clinician, pediatric
Examples of Images
Let’s take a look at what images with high or low engagement look like.
Successful Images and Associations
Here are examples of a few images and their respective associations that were linked with high engagement:
Cinematic association
Legendary association
Treaty association
Less successful Images and Associations
Now, let’s take a look at a few images that were not successful:
Vaccination association
Unfortunate association
Medical staff association
Engagement Prediction using Activation Maps: The power of the human touch?
Activation maps
Additionally, we used so-called class activation maps to show what parts of the image work. Class activation mapping is meant to visualize what a Convolutional Neural Network (CNN) is looking at to predict something.
In our class, it predicts the pixels with the strongest positive impact on engagement, i.e. what’s on the image that predicts whether a post is getting more likes, comments, and shares. [Reference] What does this mean here? That a lot of impact (red area) comes from a part of the image, with human touch.
Thus, future posts can contain more instances of touch since it is associated with warmth and positivity and can help boost engagement. As demonstrated here, activation maps are another useful social media optimization tool.
Stylistic factors: What impact do punctuations etc. have on engagement?
Just as we analyzed which words and images led to the most engagement, we considered which stylistic choices were linked with success.
Correlations of stylistic choices appearing in successful posts (represented in yellow bars) and less successful posts (gray bars)
In general, the successful posts were associated with using:
- Question marks
- Hashtags
- Common words
While the less successful posts were correlated with:
- Longer words
- Longer posts
- Using numbers
Positive vs. Negative Thoughts
After seeing all of the data, it is time to return to our initial question: for UNICEF to achieve higher social media success, should positive or negative associations be at the forefront of their SMO strategy?
To answer that, here is an overview of all the findings from the baseline analysis:
From this comparison, you can clearly see that there is a sharp divide between the associations that we find in more successful posts and those that we find in less successful posts.
The successful associations, for text and images, are more positive and instill a sense of hope. By contrast, the low engagement text and images give a sense of despair. Perhaps the darkness of the post is driving people away, as seen by the low engagement. While the warmer associations may boost engagement, leading to more success and time spent interacting with the posts.
Conclusion: Hope helps.
On UNICEF’s Facebook channel, hope associations were clearly connected to higher engagement compared to despair associations.
Using question marks, more hashtags and simple language is also effective.
Through the use of
social media impact optimization, that uses AI and machine learning, we identified which associations and stylistic choices accomplish that.
Using these insights, it is possible to be even more precise and identify exactly when to use these associations, with the ultimate results being consistently high user engagement.
Do you want to optimize your own content? Take a look at our
impact optimization service or try our online content testing and creation software
here and start your
free trial today.
More Specific Analysis of Facebook Posts
While our base analysis shows the general associations of more successful and less successful posts, a closer look shows that the 3 thematic clusters we identified at the beginning have their own associations.
Top 3 clusters identified by our embedding models based on top words
You will see that for each cluster, there are different sets of associations for text and images that are associated with success. Such detailed insights are like a gold mine for content marketers because it shows them exactly which psychological associations to tap in order to boost engagement.
