Search

The Anatomy of An Apology

If you were punched in the face on global television, what sort of apology would you want to hear?

By now we have all heard about Will Smith’s slapping of Chris Rock at the 94th Academy Awards. When it happened, opinions were divided. However, as time progressed, Will Smith, one of the most liked personalities in the world, found himself not really being liked anymore.


But why? What can data science and behavioural science tell us about how people responded to the slap - and the Oscar speech apology?


We analysed the factors attributing to Will’s negative public perceptions by collecting over 1 million tweets about the actor. We found that while having an opportunity to apologise for his behaviour during his Oscar speech, Will was neither safe nor sorry, causing his speech to be perceived as insincere and inauthentic. In this post we discuss what makes a perfect public apology and why a simple ‘sorry’ seems to be the hardest word.


 

EXHIBIT A: Public Sentiment Analysis



Using the Twitter API, we scraped public tweets about Will Smith between the 28th of March and the 4th of April. Overall, 1,086,801 tweets were collected. Valence Aware Dictionary and sEntiment Reasoner (VADER; a rule-based sentiment analysis tool) was then used to analyse the sentiment of tweets. Results show that at the start (28th of March) the general sentiment was still relatively equally positive and negative, whereas after the 29th of March, negative sentiment in tweets containing Will Smith's name became dominant. A potential explanation is that the initial shock and even amusement were taken over by more in-depth conversations about what the slap ‘really means’, which caused negative emotions such as sadness and anger. However, sentiment can only tell us so much, so the next step was to extract the themes that arose within tweets.


 

EXHIBIT B: Topics in Tweets




We used an unsupervised classification of tweets called Topic Modelling in order to find the best way to summarise tweets, and two topics were deemed optimal. One described the general ‘anger’ caused by the behaviour of both men and the other topic contained a less aggressive, but nonetheless ‘disappointed’ view of the events.


The first topic contains words such as ‘inexcusable’, ‘resign’ and ‘farce’. Opinions within these tweets range from anger over Will’s inexcusable behaviour and reactions to his resignation from the Academy, to comments about resurfaced actions of both actors (e.g., ‘1991 video resurfaces showing Will Smith mocking a bald man's hair loss…’, and Chris Rock’s previous jokes), closely followed by anger with the Oscars’ immediate inactivity and calling the whole event a ‘farce’. The second topic contained less emotionally charged tweets that mostly reflected disappointment with Will and Jada. Will Smith was criticised for suffering nearly no consequences and for displaying anti-feminist behaviour, whereas Jada was perceived as ‘manipulative’. Additionally, there was a smaller number of tweets expressing disappointment with Chris Rock for attacking someone’s medical condition. This means that the negative sentiment in tweets was, indeed, based on negative perceptions about Will and in lesser cases, about Rock and the Oscars.


 

EXHIBIT C: The ‘Apology’



Our analysis showed that Will Smith was perceived mostly negatively after the incident. However, he has apologised, hasn’t he?


We took a psychological look at his Oscars speech - using principles of statement analysis - and found several reasons why his apology was not ‘good enough’ after all.



  • Firstly, he does not really apologise at all (rather he ‘wants to’). In fact, what comes close to apology does not occur until halfway through the speech and it is not directed at Chris Rock.

  • His emotional state seemed to be very troubling, as he portrays himself as being abused, under attack and having to protect others.

  • He talks a lot about family, which seems to be very much on his mind and he makes the loaded statement about his mother at the end. This could perhaps explain why he is stressed.

  • Most importantly, instead of addressing the issue on its head, he seeks to diffuse responsibility and finds excuses by describing himself in passive terms.


All in all, his apology can be described as incomplete and insincere, which could have been the contributing factor in his increasingly negative publicity.


 

How to do it right?



We are all sometimes in a situation when we need to apologise to someone, whether as a company, brand or an individual who made a mistake. So what can we learn from Will Smith's situation and how can we apply this to our own communication?


We should not diffuse the responsibility. However, some people do this subconsciously, by using the passive voice and the third person plural, as ways of justifying themselves. To counteract this, it is worth remembering that the most effective pronouns in an apology are ‘I’ and ‘we’ (if this is written by an institution or a company).


A simple ‘I am sorry’ can be more powerful than using qualifiers (e.g. ‘sincerely’, ‘really', ‘terribly’... sorry). The assumption is that the apology is sincere, so explicitly stating this emphasis raises suspicion and weakens the statement.


An apology should be just that - an apology. Using words such as ‘I want to apologise’ or ‘I would like to apologise’ can fall short in authenticity as these apologies lack the directiveness. An apology also does not have to be long. It is, in fact, better if we do not ‘distance’ ourselves from the words ‘I am sorry’, using blame shifting, unnecessary explanations and descriptions. The words: ‘I am sorry’ are, in their simplicity, the most effective apology of them all.




At Kubik Intelligence we identify the best way forward for many companies, based on the applied data science analysis of customer data, extracting psychological drivers, motivators and secrets of your customers. If you would like us to chat about your customer data, and how we can help your business grow, please get in touch.

Unlike Will Smith, you will not be sorry!