More reasons to eat chocolate

Even talking about chocolate can make people happy. 46% of tweets about chocolate brands in the UK express positive emotions: happy, excellent, great, #fridayfeeling, lovely, smiley face, etc.

And there are more reasons why people are cheering for chocolate than just enjoyment, elation and excitement triggered by the chemicals phenylethylamine and trytophan found in it. Natural language processing and text link analysis of the tweets let me discover these reasons*.

The three brands which were tweeted about most positively in the last month were Lindt, Mars and artisan British chocolate Seed and Bean.

Lindt was making people happy with its variety of flavours. Always a bonus: 70% dark, Raspberry dark and Sea Salt dark are my favourite!

But like I’ve seen with other categories, brand social purpose is a strong driver of positive feelings towards brands. Even more so for chocolate brands, where choice is abundant and its becoming more difficult to differentiate.

Thus, the reason why people were so happy about Mars was the chocolate maker’s  announcement it is investing $1 billion to fight climate change.

The majority of positive tweets about chocolate brands were dedicated to Seed and Bean. How was it making people happy?

Being ethical tops the list with 61% of all tweets. Rich variety of flavours is second with 25% – I am enjoying Raspberry and Coconut as I’m writing this. Offering vegan options and great taste complete the list with 9% and 5%  respectively.

For Seed and Bean, and Lindt making people happy also means that they are more likely to want to try the chocolate – a link to the healthy bottom line (for the brands’ P&Ls, maybe not so much for chocolate enthusiasts).

Positive emotions about chocolate brands account for 64% contribution to people wanting to try the chocolate.

Being ethical has a 24% direct and indirect link to the desire to try. Variety and having vegan options also make a contribution of 7% and 5% respectively.

Having social purpose and generating positive emotions is paying off for a smaller and relatively less established Seed and Bean, which accounts for 68% of tweets by people wanting to try a chocolate brand.

* The analysis was performed using IBM SPSS Modeler Text Analytics.

What social data can tell us about social purpose

Increasingly brands want to know whether people think about social purpose when shopping, whether it lifts their moods or if they talk to others about it. Sometimes asking people about it in surveys may not yield accurate results: people often don’t register what they are thinking about, and may not realise that social purpose matters to them. Others may virtue signal, and exaggerate the importance of brand social purpose to them. And lastly, social purpose can have a different effect on different groups of people.

Thankfully, whilst still not perfect, actual data of conversations on social media can help understand whether social purpose matters to people engaging with different brands, and what effect it has on them.

I looked at over 14,000 tweets in the last 30 days, mentioning UK supermarkets: Aldi, ASDA, Lidl, Morrisons, M&S, Sainsbury’s, Tesco and Waitrose.

Using IBM SPSS Modeler Text Analytics package, I extracted over 5,000 concepts from these tweets and organised them into 100 themes/categories, ranging from food, drink, taste, price, customer service to emotions, employment desirability and social purpose.

When tweeting about Social Purpose in the context of UK supermarkets, people mentioned 15 different types of action, including fundraising for/donating to charities, paying fairly to staff and suppliers, improving communities and lives, helping victims of disaster emergencies, staff volunteering for charities and local communities, protecting the environment, ensuring sustainable food supplies and supporting food banks.

Present in 5% of all tweets, social purpose ranks joint 5th amongst key themes, behind food, other products, and location but on-par with customer service and offers, and ahead of wine (!), problems and online shopping.

1. Food: 11% of all tweets

2. Products: 10%

3. Location: 8%

4. Store: 7%

5. Customer Service: 5%

5. Offers, Promos, Competitions: 5%

5. Social Purpose: 5%

8. Wine, beer other drinks: 4%

9. Problems: 3%

9. Staff: 3%

9. Online Shopping: 3%

10. Plastic Bags: 2%

Text link analysis showed that social purpose was most likely to appear in tweets about Tesco, Waitrose, Sainsbury’s and M&S. It was also linked to other themes, e.g. staff, produced in the UK, plastic bags, etc.

 

Waitrose punches above its weight when it comes to being mentioned in tweets with a social purpose theme. It was mentioned in 11% of all tweets but 25% of social purpose tweets, giving it an index of 233. Tesco’s was mentioned in a whopping 41% of social purpose tweets, although a third of those were negative, accusing the retailer of pocketing plastic bag charges. M&S is underperforming relative to its overall tweets share, giving it an index below 100.

 

For Waitrose and people tweeting about it social purpose is hugely important, present in 13% of tweets and ranking 3rd only behind food and store environment.

Text link analysis shows that social purpose is interlinked with other themes. Most importantly it shows that social purpose makes people feel and tweet positively about Waitrose. It also makes them want to work for it.

Overall, premium and mid-market food retailers have higher social purpose indexes vs those competing on price. One could argue that this could be explained by their customers’ greater levels of interest in social purpose.

However, most likely this is due to the brands’ own levels of engagement in social purpose activities, how consistently they live this purpose and how they communicate it to the outside world. This would also explain variations within each category: Waitrose vs. M&S, or Tesco vs. Sainsbury’s. And with one-third of regular Waitrose customers also shopping in Lidl (Kantar/TGI), consumer segments are now shared, and there’s nothing stopping brand getting behind social purpose.

How can charities ACE it on twitter?

 

I wanted to understand how different UK charities are perceived on twitter. I chose 10 charities to start with: Amnesty International, Breast Cancer Now, British Heart Foundation, Greenpeace, Mind, Oxfam, Save the Children, Stonewall, Unicef and WWF.

I carried out thematic analysis to extract 40 keywords from over 20,000 tweets either sent by the charities or tweeted about them in the last month. I also put the tweets through IBM Watson’s tone analyser to understand how emotional the tweets were. Based on keyword meanings, associations and emotional tone of the tweets they were from, I then classified the keywords into 9 themes.

Themes and Keywords

For example, “money”, “give”, “fundraise”, “donate” keywords make up the “Give” theme. The more combative “Act” theme includes keywords “stop”, “end”, “change”, “act”, “must”, “rescue”, “combat” and “defend”. And highly emotional “Joy” theme is formed from “love”, “happy”, “good”, “trust” and “amazing”.

Give, Support and Act themes are present in tweets, which tend to be more factual, corporate and less emotional, so I colour-coded them together to form “ASK” grouping. We then move to the pink group of “Hope/Belief” and “Togetherness”, which are about “ENCOURAGEMENT”, and finally the orange “CELEBRATION” group is highly emotional and includes the themes of “Achievement”, “Gratitude”, “Pride” and “Joy”. The size of the bars on the chart above reflects the number of tweets. And as you can see, half of the tweets were the un-emotional “ASK”s.

I then used correspondence analysis to map my 10 charities against the 9 themes and ACE (Ask, Celebration, Encouragement) construct they form.

The more emotional the tweets are, the closer they are to the outer circle of Encourage or Celebrate. Those focused around “Give”, “Support” or “Act” themes are within the inner circle of Ask. The charities are also placed close to a theme that tweets from and about them correspond with most. For example, Amnesty International related tweets are calling to Act but are factual and un-emotional. Stonewall related tweets also call to Act but are Joyful (54% on IBM Watson’s tone analyser). British Heart Foundation related tweets absolutely ace it with Gratitude and high-degree of Joy (70% on IBM Watson’s tone analyser), whilst not forgetting the Give (mostly, fundraising).

Wordle of tweets from or about British Heart Foundation

Emotional tone of tweets from or about British Heart Foundation

Social media is an informal medium, and being emotional and warm is key to connecting with people, and growing a supporter-base. Numerous studies show that emotional communications are much more effective than purely rational for brand building and growth.

Depending on a charity’s objective and where it wants to be positioned, it needs to follow these ACE guidelines:

    1. Don’t just Ask, give. Give insight and stories about topics, rather than just highlighting them and quoting statistics; give insight into the inner workings of your charity and personal stories of your team. Respond to people, retweet, strike conversations.
    2. Be human, rather than corporate: use human language and don’t be afraid of showing emotions. Celebrate achievements and show pride and gratitude. Or be angry, Oxfam successfully combines Anger (60% on IBM Watson’s tone analyser) and Joy (63%), whilst Sadness (52%) combined with the Encouragement of Hope works for Breast Cancer Now, when they Ask to Support them.
    3. Encourage people by providing hope and bring them together through shared beliefs and a sense of community.

     

    Charities do ACE work – thank you!