The difference between steak and sizzle
Swoop’s Sim Simeonov sat down with MM+M’s Steve Madden to discuss the marketing capabilities of machine learning and artificial intelligence for a sponsored MM+M podcast.
Machine learning and artificial intelligence offer an unprecedented opportunity for pharmaceutical marketing to deliver differentiated messaging to patient and HCP audiences. In this “salty” episode, Simeonov explained how Swoop was able to deliver highly scalable, privacy safe information directly to consumers as well as HCPs. But first, listeners were grounded with definitions.
“AI means ‘machines doing things that we consider smart,” said Simeonov. “For example, a chatbot can now maintain a pretty good, high quality customer service conversation and solve basic problems. That was not possible one year ago.”
“Machine learning, he continued, means machines developing the ability to solve problems through experience which is traditionally done by giving it a specific set of instructions. But that data could also be historical data.” Simeonov illustrated what he meant by using the concept of weather prediction. “A machine could predict the chance of rain by looking at historical weather data, or it could learn through experience, like its robot arm failing a million times to pick up a ping-pong ball, then getting it right on the millionth-and-first try.”
With these “buzzwords” defined, Simeonov explained their significance to Swoop’s architecture as it applies to privacy-safe, pharmaceutical medical marketing.
“Machine learning and AI have become marketing buzzwords that get attached to absolutely everything, but messaging and reality aren’t always the same thing,” explained Simeonov. By way of example, Simeonov referenced a famous study in which a ML model was trained to predict the dosing of a blood thinner using different levels of privacy. At high levels of privacy, the machine could not distinguish between the drug and the patient.
“To the machine, the input was all numbers,” said Simeonov. “This traditional approach to privacy leaves you with over-sanitized data which represents another universe and makes no sense in ours.”
The difference at Swoop, is that a federated privacy architecture is built at the platform level, with privacy as the basis for everything that comes after, rather than it being an ad-on, sprinkled into the mix after the fact.
“The way companies that organize around analytics measurement typically approach data science is the way you use salt in a recipe — you add it to the end and it improves the flavor; but salt cannot fundamentally fix a flawed recipe or transform a pie into a stew,” said Simeonov. “You can’t just take a company that does ‘X,’ hire a couple of data scientists and say ‘Oh, now we’re going to do ‘X’ plus machine learning and AI…’ Or you can, but it isn’t really going to generate the highest possible quality machine learning and AI.”
As an organization built from the ground up for ML and AI, Swoop has the ability to help medical marketers target their ideal patient by creating individualized segments to reach them using both privacy and accuracy.
“In order to separate hype from reality, the simple thing to do is look at whether ML and AI is core to a vendor’s business,” suggests Simeonov. “That’s the difference between steak and sizzle.”