At Sitecore Symposium this year, we learned that every aspect of society will be turned into a prediction problem that will be solved by machine learning.
Kenneth Cukier, Senior Editor for Data and Digital at The Economist, shared this insight during the first keynote, and the buzz around machine learning continued to hum throughout the conference.
Machine Learning at its Core
Cukier helped simplify the concept. Machine learning is a branch of Artificial Intelligence, which itself is a branch of computer science.
In traditional computing, we instruct machines what to do. The machines in turn produce exact outcomes. Input to output.
In machine learning, we use historical data to teach machines to solve problems on their own. We then allow the machines to produce patterns and ultimately predict the solution. The intent is to improve the accuracy of prediction as we feed the machine more data.
While we (as in humans) are providing machines with the data and using algorithms, the machines are getting better than us at solving problems, particularly in cases where the outcome is not obvious. This has been proven in the medical field, for instance, where there are cases of machines diagnosing illnesses with greater accuracy than doctors.
Ideally, we will work alongside computers, powered by machine learning, to accomplish much more than we can on our own.
So, What Does Machine Learning Mean for Marketers?
Machine learning can be applied to many facets of marketing, specifically around targeting and predicting customer behavior.
Here are a few applications that were discussed at the symposium:
- Refine personalization rules by feeding Sitecore xDB data into Azure Machine Learning Studio – a cloud-based predictive analytics service, designed to be approachable for non-data scientists (though knowing R or Python would give you more room to experiment)
- Proactively recommend search results on a website or knowledge management platform by using an AI-powered search engine, such as Coveo
- Predict a prospect’s next move by using Sitecore Cortex – a new machine learning service on its way to the Sitecore Experience Cloud – to analyze user behavior and identify high-value actions
Quick Tips on Where To Begin:
- Reframe the challenges you are facing in your digital marketing efforts into prediction problems and business statements– Instead of, “Which channels should I use to distribute my current campaign?” ask “Which channels do I think will get the highest-level of engagement with this campaign?”
- Identify the type of data you need to predict the answer – Is the data being captured already or do you need to track it? Which system, platform or channel is it coming from?
- Work closely with the teams who own these various data sources and establish a data governance plan to keep your information clean, fresh, compliant and securely stored
While gathering a solid, vast data-set is only the starting point for applying machine learning, it is critical step that should set you up for success.
Sitecore Symposium helped validate that we are at the cusp of a major transformation within the current age of big data. Now is the time to lay the foundation and prepare for the data scientists and machine learning services to get your powerful predictions up and running.