Sim Simeonov on the Role of ML/AI in Precision Healthcare Marketing

Artificial intelligence and machine learning have become the ultimate buzzwords in pharmaceutical marketing. In many instances, companies claim to have AI/ML capabilities as a service offering to keep up with competitors and attract interest when reality is something far different. This makes it difficult for clients to distinguish which firms have AI/ML embedded into the core of their organizational structure and which are treating the technology like salt, so to speak, and merely sprinkling it on top of their business. Swoop’s Co-Founder and Chief Technology Officer, Sim Simeonov recently spoke with Steve Madden, editor of MM&M, on the publication’s podcast about how medical marketers can leverage real AI/ML companies to generate better return and reach privacy safe precision DTC audiences and identify highly valuable HCP influencers.

Defining and Distinguishing AI

The conversation began with Sim broadly defining AI, conceptualizing it as three concentric circles. He explained that machine learning is a subset of AI, and deep learning falls under machine learning’s umbrella. In business terms, AI is a machine engaging in activities that humans consider smart, and although the boundary continues to evolve, a customer service chatbot that can troubleshoot without sounding overtly robotic is a prime example.

In machine learning, problems are solved based on historical data, which is data gathered through experience or over a period of time – think of predicting the weather or a robotic arm that has failed to pick up a ball one million times but gets it right on the one million and first try after learning from all previous mistakes. This is very different from the traditional way of getting a computer to solve a problem through an algorithm, which implies that the developer knows the answer, and instead emphasizes data as the solution. A step further, deep learning is a type of machine learning that is modeled on neurological systems in the brain, with the idea that machines will soon imitate human thought processes.

While there are many ways pharmaceutical marketers can use AI/ML and, by extension, deep learning to their advantage, these platforms are generally used to refine target audiences. There is tremendous value in capturing the appropriate audience for marketers, because delivering a message to the right people serves as the industry’s ultimate objective. Since this new technology has such a ubiquitous appeal and isn’t perfectly understood, companies are shifting their marketing materials to mention these “buzzword compliant” phrases, even though their offering is still limited to counting and analysis. Sim analogizes this to adding salt and expecting it to transform a meal or conceal an inherent flaw.  

Defying Traditional Audience Activation

The prevailing process for clients seeking a patient audience is to go to a data management platform (like LiveRamp) and buy a provider’s existing, undifferentiated list—diabetes sufferers, for example. This list gets funneled into the marketer’s demand-side platform (DSP) before an ad is released. However, although a list may account for the type of condition, specificity is lost. There are no markers to indicate the stage of treatment or disorder nor other key factors such as obesity index or even gender. Critically, there is no way to ensure the members of the list will respond to the campaign, which always varies while the list remains stagnant.

“Why is it logical for there to be one segment for diabetics if there are so many different campaign objectives and so many ideal patient audiences that different drugs target?” Sim questioned, noting that it doesn’t make sense. Further adding to the illogical nature of this traditional approach is that although a company may buy a one-size-fits-all segment, success is measured based on specificity. “So, why not target what you’re measuring?” Sim charged. Simply, it’s because this approach doesn’t fit the business model of vendors — the same offenders who use AI/ML like salt.

Built by AI/ML

Because of this, clients should be discerning and seek out a business that is built on years of AI/ML experience and has created a privacy-safe data environment, like Swoop. While there’s a false dichotomy that you have to compromise quality for the sake of privacy, de-identification can be designed and built into operations; this is the best way to ensure data is HIPAA compliant without being overly sanitized to the point that results are unusable. Of course, guaranteeing quality and privacy isn’t cheap, but Swoop made the multi-million-dollar investment in our platform from the onset.

Without embedded AI/ML capability, most vendors tap out at 200-300 prebuilt segments to choose from; by contrast, Swoop has built 1,442 totally custom segments for its clients. Sim explains that our process is to ask our customers what their ideal segment is along with what the campaign hopes to achieve – we can afford to do this at scale with no exorbitant upfront cost and with a turnaround time of about a week.

Though it is more complex, this same concept can be employed to measure HCP influence. That begins with first understanding the multiple dimensions of HCP influence. HCPs can be influential because they refer patients, engage in clinical trials, write prescriptions, have a social media following, or are published – all qualify yet represent different spheres. While this is a challenge to optimize, Swoop has achieved it.

The key difference between an integrated AI/ML company, such as Swoop, is that we begin better. We maximize media budgets by building a totally custom privacy-safe audience tailored to customers versus relying on a stoic list based around a mere notion. Another point for marketers to consider when selecting a vendor is the scarcity of talent; the best data scientists want to work at companies where AI/ML is core to the business. Ultimately, whenever AI/ML are touted as selling points, clients must have a rule of thumb to separate hype from reality by investigating the company’s history and foundation.

Learn how Swoop has created almost 1500 privacy safe DTC segments. Want to find out how Swoop can help your next campaign? Let’s talk.

 

 

Comment Form