Big Data and Machine Learning based Marketing Analytics
The world generated 1.8 trillion gigabytes (approx. 1.8 x 1018 bytes) of data in 2011. Thats more bytes than the number of seconds since the universe started with the Big Bang (approx. 4.32 x 1017 seconds). Even more staggering, this data store is doubling every two years. New tools (Hadoop, HBase, to name a few.), skills (data science), and architectures (MapReduce, in-memory databases, etc.) are developing to store, manage, and analyze this ever expanding tidal wave of data.
Big Data is "high volume, high velocity, and/or high variety information assets that require new forms of processing to enable enhanced decision making, insight discovery and process optimization" as defined by Gartner. Any company that has access to large amounts of data in a variety of formats growing quickly can use the Big Data toolset to generate actionable insights that increase revenue, cut costs and/or increase customer satisfaction. For example, AirBnB identifies its most valuable users by loading its social media data into a Hadoop cluster and figuring out which users have the most influence on their social networks for recommending AirBnB and generating revenue. It targets marketing initiatives at these influencers to maximize revenue.
Big Data applications often use Machine Learning techniques to create predictive models from analysis of past data.
For example, Netflix suggests movies you might like to watch next based on movies watched by you and others in the past. Machine Learning (ML) is a sub-branch of artificial intelligence focused on building systems that learn from data. Within ML, Supervised Learning builds a model from data where the answers are known and then uses the "learnt" model to answer questions about previously unseen data. Google's GMail uses this technique to identify spam in your inbox. Unsupervised Learning analyzes data without labeled answers and tries to identify trends and patterns within it. Geophysical Insights is a company that uses unsupervised learning processes on seismic data and determines probable areas for oil and gas exploration.
IDC forecasts that the worldwide business analytics market will grow to $50.7 billion by 2016. Within that market, there is a lot of innovation happening in the marketing analytics space to help companies tie marketing spend to revenue, make ROI based marketing spend decisions, identify their most promising leads, etc.. For instance, Lattice does predictive lead scoring for marketing by tying into existing marketing automation systems (Marketo, Eloqua) and using unsupervised learning to discover patterns in leads that become deals. Once these patterns are identified, it annotates leads with a lead score that predicts lead conversion likelihood.
Another company, AgilOne, has created a marketing analytics platform that creates a) a data layer by collecting, cleansing, and integrating internal and external customer data, b) an intelligence layer using supervised and unsupervised learning procedures, and c) an action layer for marketers to convert the insights generated by the intelligence layer into actions like automated campaigns to reach customers via email, direct mail, call centers, etc.
This data to insights to action layering is a common pattern that is emerging across various industries. BrightFunnel offers a turnkey solution that operates at each of these layers aimed at Chief Marketing Officers. They can pull in marketing data from various sources, get machine learning powered revenue performance insights, allocate marketing spend across channels based on revenue predictions, and benchmark performance against self and peers.
There is a multi-billion dollar opportunity for enterprises to use marketing analytics tools to create a data driven positive feedback loop that continually increases efficiency and ROI of their marketing spend. It will be interesting to watch how this space evolves.
About Author
Nisheeth Ranjan
Co-founder & CTO-BrightFunnel
Based in San Francisco, BrightFunnel is a big data analytics company which connects the dots between marketing data silos to generate predictive, actionable revenue insights.