Data platform company Tealium is positioning itself as an AI enabler on three fronts: data preparation, using machine learning to identify audiences and, activating those insights through marketers’ digital channels.
“We are not an AI company, we are an AI enabler,” Tealium’s CEO Jeff Lunsford told Which-50 during a recent visit to Sydney.
- Leadership Webinar: Which-50’s 2019 Outlook and Business Transformation Drivers webinar is set for November 27. Register today!
Early next year Tealium is launching Tealium Think, a new product which will host machine learning models for its customers. Think will analyse data collected by Tealium to identify audience segments in real time.
For example, on an ecommerce site Tealium will take the visitor-level information from CMS and run it through Tealium Think to identify which customers have a high probability of abandoning the purchase at each stage of the conversion funnel.
The machine learning scores are then put into Tealium Audience Stream which sends instructions “up-stack” to email, mobile push or content management systems to take a certain action. For example, if the system identifies there’s a high risk a customer will abandon a purchase, marketers can program rules which will automatically offer the customer a 5 per cent discount based on the model’s output.
“Think of it as audience creation based on all the information that is available and taking action on those audiences,” Lunsford explained.
The move to support AI is a natural progression for the software company which already acts as an independent data layer underneath a company’s suite of software applications, Lunsford said.
“Every 18 months we have rolled out a new service which is complementary with our existing service. It’s all been built organically, none of it through acquisition. Tealium Think is the fifth major service,” he said.
The addition of machine learning capabilities to martech is being driven by the pressure to act in real time and the sheer volume of data that is generated through digital interactions.
“You can’t pull your customer data into a data warehouse and then look at what people were doing last week. Marketers need to target based on real time interests at the point in time when they have the highest propensity to transact,” Lunsford said.
“If you can’t do these things your business is on a long slow decline to mediocrity.”
Lunsford argued companies will have to leverage machine learning to stay competitive.
“There’s now so much of this data you need to use machine learning/AI to process it and understand it,” he said.
To do that data must be machine learning-ready. Lunsford said Tealium also plays a role in normalising, cleansing and filtering the data before it is fed to algorithms.