What is Predictive Analytics?
According to Wikipedia, following is the definition of Predictive Analytics:
The core of predictive analytics relies on capturing relationships between explanatory variables and the predicted variables from past occurrences, and exploiting them to predict the unknown outcome. It is important to note, however, that the accuracy and usability of results will depend greatly on the level of data analysis and the quality of assumptions
When it comes to predicting customer behavior, retailers (both online and offline) rely on historical data about a set of customers, and try to capture relationships between products, demographics, spends and then try to predict the outcome of a particular customer coming to the online store or walking into the retail store.
But only past transactions and clickstream and that too at an aggregate level cannot really be enough to predict what a customer might purchase. As mentioned in the Wikipedia definition above, the accuracy greatly depends on the quality of assumptions. And this assumption is flawed to begin with.
I have been a working professional, then an entrepreneur, then a student, and then an entrepreneur again. My purchase history will never tell you about my circumstances, which are prime factors driving my spending. If I went from being a student, spending only a few hundred dollars, to the CEO of a well-funded startup, my spending potential has increased multifold. This life event of mine, is not captured in any of my past transactions.
The one place where my life events get captured, is a Social Network.
Infinite Analytics’ predictive analytics use this social data about a customer, along with all the legacy data sources, like transaction history and clickstream, to better predict a customer outcome. We use a mix of predictive, descriptive and decision models for our prediction scores. The direct upshot of this is the fact that our clients are seeing close to 25% increase in conversions based on our predictive analytics.
Our Natural Language Processing (NLP), Machine Learning and Semantic Technologies help us establish relationships between users, brands, stars who endorse those brands, spending potential & even customer influence.