What machine learning and semiotics can reveal about a brand’s values
Brand values and personality define what your brand stands for and the response you want it to evoke. But guess what? Communications from your brand might not always convey the intended personality or be consistent across even a small selection of brand touchpoints.
We’re often asked to test communications and assess brand perceptions, addressing things like their relevance, distinctiveness, alignment and consistency across different touchpoints like websites, promotions, ads, products and services.
One method we use to surface the meaning conveyed in these touchpoints is semiotics.
What is semiotics?
Semiotics is the study of signs and symbols, the core tenet of which is that everything – language, images, objects, places, etc – has a meaning in culture. Semiotics aims to interpret these ‘signs’ and to decode their symbolic import to reveal both explicit (intended) and implicit (potentially unintended) meanings, and the likely impact of these meanings on human behaviour.
Until recently, however, semiotics has mainly been carried out on relatively small data sets at a painstakingly slow pace, by individuals whose skills often seem more akin to art than science.
As our culture becomes saturated with text, images and objects, it is clear that understanding the meanings they contain is as essential as ever, but that the pace and scale of semiotic analysis need to increase dramatically to reflect this new world. Here, new technology can help us, via the emerging discipline of quantitative semiotics.
Take websites and communications on the internet: every day, a plethora of new content is produced and shared in the hope of influencing audiences. Implied and explicit messages are strewn across websites. Each webpage is structured in a specific way and contains multimedia content that is laid out in unique ways on every site.
What does the sum total of your website communicate about your brand? And if your brand has a complex architecture – as so many automotive, FMCG, media and bank brands do – how do the various components communicate different, yet coherent, meanings?
Defining and communicating your brand personality
When building foundational brand platforms, defining the brand’s personality and how it should be communicated forms part of this process. ‘Archetypes’ are often used to establish the most suitable personality that should be associated with a brand. The idea is derived from Carl Jung’s theory that human beings share universal meaning structures that are rooted in the ‘collective unconscious’.
All of us can, therefore, perceive these archetypes without necessarily being aware of it. It is both the universality and the unconscious aspect of these archetypes that make them so powerful for brands.
Jung identified 12 main archetypes:
Think of your favourite movies and you’ll notice narrative archetypes in all of them. Harrison Ford is an unmistakable hero in Indiana Jones, Princess Leia cuts a perfect rebel figure in Star Wars, George Bailey in ‘It’s a Wonderful Life’ is everyone’s favourite everyman, while Dorothy in the Wizard of Oz perfectly captures the innocent.
Using semiotics to quantify brand archetypes
We partnered with Signoi, whose platform uses machine learning and semiotics to explore what brands’ websites are saying and how they speak to their nuanced audiences. The platform decodes and measures meaning from unstructured data, such as the messages conveyed by the brand’s customer proposition.
Signoi’s machine-learning models have been trained to spot and measure useful signifiers of brand and marketing concepts in language or imagery. Concepts can range from the purely descriptive through to the metaphorical – such as the Jungian archetypes. By scraping content that’s freely available (words and pictures) from the relevant websites, it compiles a rich database for analysis. The content is then parsed through the Signoi platform and automatically scored according to the presence or absence of these signifiers, allowing one brand to be reliably compared with another with no human intervention.
This allows us to directly compare one brand’s archetypes with another, using unstructured data that’s freely available. As an example, let’s look closely at a major brand – Virgin – and see what semiotics reveals about the values it communicates.
How well does Virgin communicate its brand values?
We recently ran an experiment to test out the use of quantitative semiotics to determine which archetypes a brand’s written proposition conveys on its websites. We applied this approach to the Virgin brand websites in the UK by scraping the core proposition statements from each of their operating company’s websites (e.g. Virgin Media, Virgin Money, VIrgin Atlantic, etc).
Virgin explicitly aims to position itself as a Rebel brand, spearheaded by its inimitable leader Richard Branson, who restless spirit, entrepreneurship and championing of consumer fairness. How closely does this map to how the brand presents itself in practice?
What we learned
We determined that the Virgin Group as a whole is not communicating this Rebel archetype as consistently as might have been hoped. It is closer to being an Explorer, a Creator, and a Sage than a Rebel. No bad thing – but food for thought for the brand.
Not surprisingly, these archetypal associations are completely different from those of other major brands like Tesco – with its strong emphasis on embodying the Everyman – and John Lewis – where the Caregiver and Lover are the dominant archetypes. Here is what each brand looks like when its archetypal associations are mapped out visually.
Importantly, our analysis also revealed significant discrepancies between the archetypes associated with the master brand, Virgin Group, and its sub-brands. To some extent, this is desirable. For instance, Virgin Money and Virgin Atlantic should probably be strongly associated with the Ruler, with its emphasis on mastery and competence, while Virgin Active might profitably be associated with the Caregiver or Magician, who will transform its members.
Virgin’s example goes to show that the values a brand sets out to convey and the values it actually conveys may not always align – and that it pays to be aware of these finer details in order to ensure that a brand’s positioning, content and communication are as cohesive as can be.
Quantitative semiotics can quickly reveal archetypes like these – for category analysis, brand mapping and positioning diagnosis – allowing branding and communications teams much greater control over the core messages being transmitted by advertising or visual comms material. Combined with audience insights on brand perceptions and communications – which tell you what messages people are receiving from the brand – this approach can highlight ways to ensure that brands stand out in their category, and consistently and credibly communicate their values and personality.
The post What machine learning and semiotics can reveal about a brand’s values appeared first on Econsultancy.
Source: Customer Experience