For many years we have witnessed marketing and IT working together to transform organisations into smart, fast moving, innovative businesses. Gartner’s 2018 trends highlight how Machine Learning (ML) can be used to interpret the interaction between devices, people and surroundings in a more intelligent way; but when the conversation turns to Artificial Intelligence (AI) it is often associated with sci-fi movies, not marketing and IT.
At Switch, we’ve been applying new marketing capabilities for ML using Sitecore’s Cortex*. Our approach has been to take this new Cortex technology and apply it to common marketing activities in order to support the application for personalisation, re-targeting, ad-bidding and analytical reporting. This application isn’t the ‘art’ of marketing, but it automates the tracking of customer behaviour across multiple digital touchpoints, changing the way in which analytics are captured and helping marketers to make smarter decisions without having to manually crunch huge amounts of data themselves.
It all starts with Digital Goals. In Sitecore we create goals for campaigns, form completions, page events, transactions, and even offline activities such as print. Details on how to setup Sitecore goals are located here.
A consistent problem for the marketing profession is the management and up-keep of goals and goal-weighted values. While it is easy enough for someone to set up a website goal, in order to understand the weighted value of each (while adding new goals and refining existing values) the maintenance often gets left behind. But we need it! We need goals and weighted values to ensure that personalisation, content optimisation and reporting is accurate and always aligned to the content strategy. This is where ML comes in.
To take a quick history lesson, ML, in its simplest form, is the creation of models of data which, as the database is populated over time, become smarter. One of the earliest examples of ML is Deep Blue, the first computer to win a chess game and beat the world champion chess player in 1996. Deep Blue didn’t win immediately, but instead processed and applied the data gathered during each game - gradually learning by recognising and understanding its opponent’s strategies.
How does this work in Sitecore Cortex, you ask? Imagine that your chess pieces are your goals in Sitecore, and your moves are the visitors. As people move from one goal to another, the machine learns. The goals’ value weightings can be automatically adjusted to best represent which goals lead to the desired outcomes. We simply model the goals and values based on the outcomes we want the visitor to achieve.
Currently, to do this manually requires hours of planning, execution and maintenance. The manual application of accurate goals and value weighting is questionable, and value weighting of a goal is often subjective. Not only is ML a smarter way of managing analytics, but the targeted personalisation which is triggered by a goal is more precise, increasing the likelihood for a customer to engage with the content and convert.
It’s an exciting time to be implementing AI and ML within Sitecore for our customers and the powerful results that they bring. To find out more about how your team can start implementing AI and ML today, contact us at email@example.com.
* You can view a short video about Sitecore Cortex here. There are also a number of articles online for system architects discussing how the process of modelling combined with algorithms takes place in Sitecore.
Subscribe for the latest news and events from our talented team and community network.