In the word of digital marketing, one of the more controversial moves I’ve seen recently was from U.K. car insurer Admiral. The company recently announced that it would begin offering car insurance discounts to less risky customers based on voluntarily provided social media data. The insurer would analyze Facebook likes and posts, and could analyze the language and patterns. This allows it to identify behaviour and personality traits which predict a higher or lower risk compared with the average for that demographic profile.
While Admiral’s plans were eventually scrapped due to Facebook’s data privacy policies, the simple truth is that many digital footprints are already being harnessed, analysed and shared to assist digital marketing efforts (from consumer goods to political parties). It’s possible to increase conversions and reduce the cost of acquisition through understanding digital visitors better and ensuring that adverts reach the “right kind” of consumer.
Did you know your browsing behaviour is already being tracked and traded?
Trading of such data between interested third parties is also growing, with data management platforms (DMPs) providing digital marketing teams with behavioral profiles based digital data collected across a wide variety of websites over a period of time. For example, an online clothing retailer may identify you, a first-time shopper on the site, and query this centralised DMP to get a detailed shopper profile indicating your likely segment and product category interest probabilities. This is based on your previous visits to sites such as airline and hotel bookings, the news articles you've read and your cellphone company.
Are consumers willing to do this?
Generally, consumers seem happy to share much of their digital data (even if they're not aware of the full extent of how their data is used) as long as they perceive some value – particularly when it comes to personalisation of the website experience. A variety of research showed that well over 50% of consumers expect brands to know them more intimately, 90% are willing to share basic browsing behaviour, and over 60% are willing to share detailed personal information in order to receive a more personalised experience.
Facebook knows you better than your own spouse …
Recent behavioural analysis research at Stanford University aimed to predict a customer’s likely personality traits, beliefs and opinions based on an analysis of the digital behaviour of 8 million social media accounts (compared with online psychometric tests). The algorithms showed that personality traits such as generation, introversion/extroversion, family situation, race, smoking/alcohol use and political or religious beliefs can be predicted to a high degree of accuracy based on an analysis of an average of 200 Facebook likes per user. The researchers believe that, if given over 300 likes to analyze, the algorithm could predict the subject’s true personality better than the his or her own spouse.
The exciting part of such research from a marketer’s perspective is the ability for big data analytics, coupled with machine learning techniques, to predict a consumer’s likely personality traits (which would otherwise have to be extracted through direct questioning or surveys) through an analysis of a social media profile.
Advertising based on personality or behaviour can be powerful – for example, a Facebook ad for for exactly the same cosmetics brand can be tailored to introverts and extroverts differently. One advert could be worded in such a way as to offer extroverts the promise to help them to stand out from the crowd, and the other could offer introverts an internal feeling of confidence.
Harnessing the existing rich behavioural data from your website
Even without access to a consumer’s social media likes, digital marketing teams have access to vast amounts of digital behavioural data from their owned media, such as the “clickstream” data from their website and mobile app. This provides rich data to inform site personalisation. For example, a mobile operator can quickly identify a likely affinity for Samsung vs. Apple. They can then ensure that Samsung-related offers are presented more regularly … thereby boosting conversion rates and simplifying the user’s experience.
With the ever increasing percentage of the customer journey taking place through digital channels (and often across multiple devices), marketers need to identify that unique individual across devices and channels. It’s then vital that they integrate this clickstream data with the rest of the customer profile (much of which is extracted from non-digital channels such as stores/branches, billing/POS systems and customer relationship management systems.
The key is to link digital and traditional data into a single customer behavioural profile
By linking such online and offline data together, a much richer profile of the customers behaviour can be created and used to perform more advanced predictions through analytical models. Exhibited digital behaviour can be a strong indicator of likely product purchase propensity or brand affinity. This can be used to further personalise digital marketing campaigns and offers across both the initial digital channels as well as in the call centre or branch, as well to inform direct outbound campaigns via email and SMS.
For more information on how to harness both online and offline data for richer personalization of the customer experience, please read the white paper: Using SAS to Deliver Analytically Injected Digital Personalization for Online and Offline Data, or contact me at email@example.com for more details.
More information on Stanford’s MyPersonality project can be found in this insightful edition of the Blindspot podcast with Dr. Michael Kosinsky.