I presented 108 keynotes in 2015 when I worked at Salesforce. Earlier this year I jumped at the opportunity to work at IBM talking about Watson, and have since given 42 keynotes around the world, talking about marketing technology, trends in customer behaviour and the ethics and opportunities around AI and cognitive computing. Much of that time has been spent talking to executives about the value of IBM’s own cognitive computing system called Watson, and how he can analyse not just interactions, events and transactions of individuals, but the emotions, sentiment and behavioural analytics of audiences. Analysing emotions is hard, but the fact that Watson can do it faster and more accurately than any other computing system, makes it a fun technology to play with. It’s one of the reasons I joined IBM.
So I thought I would put Watson to the test…
“Marketing is no longer about the stuff you make but the stories you tell”. Seth Godin
In the short time I have been at IBM I have witnessed various engineers and IBM data scientists playing with Watson’s API’s in order to do all kinds of wonderful things. A few of them are:
Suffice to say that IBM has its fair share of emotional stories to tell.
I like the idea of having fun with technology, because it’s often in those playful moments that break through innovations are made. So it got me thinking ~ if a technology like Watson can help to analyse and create the perfect [dress / song / movie / piece of marketing content] AND can read over ten million records a second to help oncology consultants make more accurate diagnoses, then surely he can shed some light on my stage-dwelling ramblings? Could Watson analysis suggest the shape of the perfect keynote? Tell me what worked, what didn’t and why?
Nothing to lose, I thought. So I set to work, putting all of the presentations that I have performed over the last 6 months, asking Watson to analyse them for emotional context (using natural language processing and personality insights) – in order to attempt to find any correlations between the good keynotes, the average ones and the ugly ones. And yes there were a few ugly ones.
The results were staggering.
Not unlike the wonderful TED talk that Nancy Duarte gave a couple of years ago about the shape of the perfect presentation, it turns out that the shape of all my keynotes which received the highest audience scores, also had exactly the same shape!
Of course this small “eureka” moment means a lot more to me that it will to anyone else, but it reminds me that this technology is real – it’s not smoke and mirrors (there is no magical wizard hiding behind a curtain trying to make the demos work)! It’s technology that can help anyone solve problems, whether they are trying to cure diseases, send their customers the right email, or just understand why some of their presentations resonate with audiences more than others. I have been using a similar format for a couple of years ever since I learned that Steve Jobs used the same narrative format for his keynotes that Pixar used for all of their films.
So what is this magical shape?
Well it’s quite simple – it’s based upon the shape of a story that Nancy Duarte introduced me to a while ago, but I have refined it to include various emotional triggers appearing at specific parts of each presentation. I have reams of data that I won’t bore you with here, but I drew the shape of my successful keynotes as I was trying to interpret that data, based upon the personality insights API that Watson used to analyse all of my presentations.
It’s all very interesting stuff, but what was even more interesting to me was the reason why the less successful presentations weren’t as effective. Here’s four of the biggest insights I discovered from the analysis.
Watson understood the emotional intent behind each of my highest performing slides (based upon feedback from audience surveys or feedback I noted after each presentation) – and attributed certain ‘feelings’ to each slide, recommended the type of emotions that certain slides should trigger.
Why are emotions so important? Aside from the obvious, many public-speaking coaches actually play down the importance of language and suggest emotional context is far more important. they often cite research published in 1967 by Professor Albert Mehrabian who claimed that only 7% of the effectiveness of communication was down to language, white 38% was directly linked to the tone of voice, and 55% from body language.
This is why I thought it would be really valuable to analyse the emotions behind my content, using IBM Watson’s Personality Insights engine. Analysing content from any source, it creates starburst diagrams like this, to help visualise the most dominant emotions.
The Most Effective Emotional Triggers for Each Slide
Translating those emotions at certain points of my “perfect” presentation (which it turns out should be no longer than 30 minutes, means that all of my presentations should fit into this template.
So, let’s compare what one of my old presentations looked like, compared to the new Watson “Perfect Keynote” model….
(One of my first IBM “official” keynote decks)
(My Madrid keynote from earlier this week “built” by Watson)
They look quite different don’t they?!
All of this just because of a simple story shape transforms the way that people receive, digest and understand information. This is why scientists, anthropologists and biologists like Prof. Brian Cox, Simon Sinek and David Attenborough are such successful speakers. They break down complex topics into stories that the average person on the street can understand.
“People are persuaded not by what you say, but by what they understand”. John Maxwell.
So what do my scribbles actually mean?
I could have created a sexy dataviz or extracted all of the data into Excel or Tableau but where would the fun in that be when you can draw it with coloured pencils, over a Rioja in the Ritz hotel in Madrid. So that’s what I did. I sat in the bar trying to make sense of all the data while I was taking my time drawing various shapes on graph paper. What you see above is the shape that best represents my perfect keynote. As with all analysis like this, what works for me probably won’t work for you, but that’s a good job. Because if it was that easy I’d probably be out of a job.
Breaking it down, these are my findings from Watson’s data about how I should structure my perfect keynote:
And to think, all of these insights came from a dumb computer who just ingested a ton of data, analysed it, understood my emotional intent, measured it against previous language of a similar substance, and then created some hypothesis about what to do with it. That my friends is the power of cognitive computing.
It was an incredibly arduous exercise as my data was raw and unstructured, but Watson’s findings have left me with a much better understanding of the style of presenting that suits me best. From this I have taken eight key takeaways that you may use of discard as you see fit.
8 Takeaways For My Perfect Keynote
Fascinating stuff isn’t it?
You can learn more about the kind of technologies I’ve been playing with at IBM.com/cognitive. And if you’d like to discuss any more of my findings behind the data, and would like an even geekier conversation, I’d be happy to oblige. Just drop me a note in the comments or say hello on twitter @jeremywaite.
“Whoever tells the best stories goes home with the most marbles”.
The Personality Insights service from IBM Watson is based on the psychology of language in combination with data analytics algorithms. The service analyzes the content that you send and returns a personality profile for the author of the input. The service infers personality characteristics based on three models:
A well-accepted theory of psychology, marketing, and other fields is that human language reflects personality, thinking style, social connections, and emotional states. The frequency with which we use certain categories of words can provide clues to these characteristics. Several researchers found that variations in word usage in writings such as blogs, essays, and tweets can predict aspects of personality (Fast & Funder, 2008; Gill et al., 2009; Golbeck et al., 2011; Hirsh & Peterson, 2009; and Yarkoni, 2010).
IBM conducted a set of studies to understand whether personality characteristics inferred from social media data can predict people’s behavior and preferences. IBM found that people with specific personality characteristics responded and re-tweeted in higher numbers in information-collection and -spreading tasks. For example, people who score high on excitement-seeking are more likely to respond, while those who score high on cautiousness are less likely to do so (Mahmud et al., 2013). Similarly, people who score high on modesty, openness, and friendliness are more likely to spread information (Lee et al., 2014).
IBM also found that people with high openness and low emotional range (neuroticism) as inferred from social media language responded more favorably (for example, by clicking an advertisement link or following an account), results that have been corroborated with survey-based, ground-truth checking. For example, targeting the top 10 percent of users in terms of high openness and low emotional range resulted in increases in click rate from 6.8 percent to 11.3 percent and in follow rate from 4.7 percent to 8.8 percent.
Multiple recent studies disclosed similar results for characteristics that were computed from social media data. One recent study with retail store data found that people who score high in orderliness, self-discipline, and cautiousness and low in immoderation are 40 percent more likely to respond to coupons than the random population. A second study found that people with specific values showed specific reading interests (Hsieh et al. 2014). For example, people with a higher self-transcendence value demonstrated an interest in reading articles about the environment, and people with a higher self-enhancement value showed an interest in reading articles about work. A third study of more than 600 Twitter users found that a person’s personality characteristics can predict their brand preference with 65 percent accuracy.
The following sections expand upon these high-level findings to describe the research and development behind the Personality Insights service. For more information about studies that apply the service to tangible scenarios, see The service in action.
For the Personality Insights service, IBM developed models to infer scores for Big Five dimensions and facets, Needs, and Values from textual information. The models reported by the service are based on research in the fields of psychology, psycholinguistics, and marketing:
For more on how Personality Insights are calculated based upon established >> http://www.ibm.com/watson/developercloud/doc/personality-insights/models.shtml#outputValues
If you enjoyed all that – you’ll LOVE this >>