For the tenth year, Med Ad News has chosen new Pharmaceutical Marketing Ventures to Watch that could change the way pharmaceutical products are marketed and sold.

 

During this past October, the Med Ad News staff began its annual search for the future of pharmaceutical marketing. We sought out young companies, spin-offs, offerings, and ventures to profile that are providing the most innovative and interesting products, services, or marketing opportunities to pharmaceutical companies and the healthcare community. This year’s two profilees are both focused on targeted digital advertising, but in very different ways; one is an AI/machine learning module that can be embedded in ads and respond to physicians’ questions, and the other is a “smart” marketplace for buying and selling digital ad inventory. Here are Med Ad News’ newest Pharmaceutical Marketing Ventures to Watch.

@Point of Care – ad-embedded
@SK with Watson tool

@Point of Care’s @SK with Watson tool can be embedded into
third-party ads and offers quick responses to physicians seeking
information; the tool’s underlying decision support module has
already been used by more than 200,000 clinicians in 29 disease
categories.

@Point of Care is bringing static digital ads to life by adding its ad-tech tool @SK with Watson. Built with artificial intelligence, cognitive, and machine learning technology, the tool ingests the product data/content and supports medical professionals seeking information by providing real-time, succinct answers to questions. Its innovative functionality, company executives say, offers healthcare marketers unsurpassed levels of granular insight via analytics on their customers.

The @Point of Care team has been through several businesses over the course of 30 years, each one building to some degree on the previous ventures. All the businesses were in the medical information and data space, from educating doctors to formulary insights to medical news. “The most recent business, which was sold to Everyday Health, was MedPage Today – breaking medical news written on the clinician level,” says Robert Stern, @Point of Care’s CEO. “This business taught us a great deal about gathering just-in-time information and turning this information around in minutes. I.e., how to gather and parse data in real-time and offer new metrics not seen before in the industry.”

So what was next? During the course of their work on MedPage Today, the team discovered that there was no resource for clinicians that could give them immediate, short, targeted answers to their point-of-care questions in the exam room without having to go through mounds of information and self-curate an answer. “In a 10-to-15 minute patient visit, that was not going to work,” Stern says. “So, we took our knowledge from MedPage Today on how to gather and update clinically relevant evidence based information on the fly and present it in a crisp pointed answer useful to the clinician while the patient was in the room, a virtual consult.”

As luck would have it, IBM introduced the Watson AI/Cognitive machine learning tool and was looking for medical applications just as the @Point of Care team was looking for an AI tool. “As we were looking to leverage our expertise to build a next generation healthcare AI platform with IBM Watson’s natural language expertise and cognitive services, we came to the partnership with deep level content and knowledge services,” Stern says. “To make a long story short, we entered into a partnership which is now more than four years old.”

In the meantime, @Point of Care’s in-house development team has built on and added its own proprietary cognitive models and tools to work with the Watson product, adding more depth and specific medical technology AI/and machine learning that leaders say make the combined tools best in class for clinical decision support. “There are several large players in the field – Elsevier, DynaMed, UpToDate, Epocrates, Medscape – but none with the game changing tech we have built,” Stern says.

So @Point of Care was able to bring all this intelligence, artificial and otherwise, to bear with its Cognitive Decision Support platform, offering physicians a way to access relevant content quickly at the point of care. But the next innovation – the one that earned @Point of Care this profile – went a step further.

“So, once we built the clinical decision-making tool and could leverage our skills and experience in developing an expertly trained cognitive data model, our CTO David Setiadi, Ph.D., and our commercial sales lead, Margo Ullmann, said, ‘Why can’t we lift the AI/Cognitive machine learning component and make it available as a standalone to supplement ads and make them come alive in real time, ads that can answer questions posed by clinicians or patients in real time?’” Stern told Med Ad News. “So, we mocked up a few demos, developed our @SK with Watson Chat Bot, and left pitch rooms fascinated and buzzing with ideas being tossed around about, how do we get this idea executed as soon as possible.”

The @SK Chat Bot is only about two months old and is already creating plenty of buzz. “As far as ad agencies and their clients go, this is a perfect, controllable product,” Stern says. “They provide the approved content, resources, and answers and we train the @SK Chat Bot to respond with those specifics or refer their question(s) to appropriate channels. So, no surprises because it’s a controlled media offering that legal and regulatory can guide.”

How does it work? Agencies develop and supply the creative and media buys, which they execute. Then, @Point of Care takes their creative and approved product information and pulls it into the @SK AI/Cognitive environment, and prepares the Q&A portion to be inserted as per the agency creative and size needs. “We can also change programming on-demand, so as labeling or other info changes we can make behind the scenes additions without taking down the advertisement,” Stern says.

Although the @ASK Chat Bot itself is relatively new, the clinical decision support tool from which it emerged is very much not; according to @Point of Care executives, it’s been used by more than 200,000 clinicians in 29 disease categories. So there’s plenty of data support – and evidence-based linkage to changes in prescriber behavior – behind the curtain.

“Our real value on the backend is insight,” Stern says. “What we can analyze on a de-identified aggregated basis is what is the clinician is looking at in terms of answers to their questions during the patient encounter that changes their prescribing or treatment or continuation with current treatment. No one else can do that.”

And of course, it’s all highly targetable. “@sk with Watson, and the @SK Chat Bot, are totally targetable, and the answers are different for patients than clinicians.” Stern says. “You can slice and dice messaging to any segment of audience that fits the client criteria.”

What’s next for the @SK Chat Bot? Pretty much any space where medical knowledge is needed. “Chat Bot for sales reps and MSLs, hospitals, physician offices … making journal articles come alive … adding to digital screens in hospital rooms to answer specific patient questions, patient support screens in waiting rooms, and disease-specific patient apps,” Stern told Med Ad News. “Plus, news articles that can go further by offering chat bot to talk to experts, medical sites for doctors offering author chats, and voice chats with various speaker products available and so on.”

The company will also be looking at predictive analytics with prepopulated patient info, with a target of the second half of 2018. After that the real long-term home run, Stern suggests, will be patients interacting with their own doctors to monitor and handle patient questions, including patient record data matched to the physician’s protocols to isolate next steps and treatment options. “Can I take Tylenol with this drug, should I come in, and Watson triages the appointment,” he says. “In the appointment calendar, the doctor gets all the latest info and insights they need before the patient comes in the office.”

What’s important, Stern notes, is that none of these tools are intended to replace clinicians; they are intended to maximize the value of the patient/physician interaction. “This is about providing information in seconds that a person would need years to parse through and giving the clinician a tool to make it relevant to each patient differently depending on history, comorbidities, and challenges,” he says.

 

Proclivity – LayerRx marketplace

Proclivity’s LayerRx is a “smart” marketplace for buyers and sellers of digital media advertising inventory in the healthcare professional space; among other things, the tool offers predictive modules that guide users towards optimal buying and selling decisions.

Proclivity’s LayerRx is a “smart” marketplace for buyers and sellers of digital media advertising inventory in the healthcare professional space.

Proclivity was founded in 2007 after its CEO, Sheldon Gilbert, deferred graduate school in computational genomics and decided instead to launch a startup – out of his bedroom – based on an algorithm and overall predictive analytics software platform that he had developed after several years of thought.

“Around the time that I was graduating from college in 1997, the human genome was being decoded and when I joined a molecular genetics lab as research assistant, I realized two important developments,” Gilbert says. “First, we were in the midst of the dotcom revolution and the world was becoming rapidly digitized and there would be an unprecedented amount of ‘data exhaust’ being generated from all of these new online businesses that could yield predictive insights. Second, as the genome was being mapped, there were so many new genes, genetic properties, and ontologies being discovered at such a rapid rate that the canonical relational databases models used to store and model this new information needed to be reconsidered because the data schemas were in constant flux. This gave rise to new approaches to data modeling such as EAV (entity-attribute-value) models as well as ontological databases which became an important factor in the rise of NoSQL databases nearly a decade later of which there are now several such as MongoDB, Cassandra, HBase, etc.”

So the epiphany for Proclivity came when Gilbert realized, after leaving the lab and joining some friends from college to launch a start up in another industry, that he could develop algorithms that would take massive volumes of unstructured website log data and in-store data to effectively predict what products consumers would buy at what price-points with a stable degree of accuracy. “Then in 2010 with the rise of programmatic advertising, we extended our platform to allow brands, seeking to buy media targeting desired customers to bid and buy ads based on the real-time predictive transactional value of these target customers,” he says.

This experience allowed the Proclivity team to gain unique insights on the first-generation of real-time advertising exchanges and discover the extraordinary levels of information asymmetry in the market between buyers and sellers. “We felt that this would limit the efficacy of these online marketplaces by limiting the transactional and trading liquidity between buyers and sellers due to the lack of expertise in how to properly utilize data to compute the value of the unit of trade (in this case the ad impression) being bought or sold,” Gilbert told Med Ad News. “We knew then that we would some day need to build a better electronic marketplace where both sides don’t just have data but, more importantly, know how best to leverage data and developed best-in-class algorithms to effectively compute optimal bid and ask values for trade.”

The opportunity to do so came in 2013 in a new industry that was largely analog and ready to become more automated: pharmaceutical marketing for healthcare professionals. “After we applied our tech platform in the retail/e-commerce sector for several years, we were approached by one of the largest media buying agencies in the pharmaceutical industry to develop the first-ever automated media buying platform to help them reach verified HCPs in real-time with the right message and right time across any digital medium,” Gilbert says. “As we built this buy-side platform, we soon realized that the media sellers (publishers) in this market didn’t have the appropriate converse technology, an automated sell-side platform, to interface and engage with this new automated buy-side platform.”

This provided Proclivity with the opportunity to formally develop LayerRx, the type of electronic marketplace that its leaders had imagined that would allow buyers and sellers to trade more effectively with a more symmetric capacity for yield optimization and with a focus on limiting/preventing data leakage.

“What we didn’t anticipate was applying it in the pharmaceutical HCP marketplace but it turned out to be ideal because it has a lot of unique features, unlike all of the ad marketplaces,” Gilbert says. “For starters, HCP media is, by far, the highest value media across any industry and there is unparalleled stringency on securing digital data related to HCPs in order to maximize the yield on a limited supply – there are roughly 1.2 million HCPs in the United States.”

What does LayerRx look like to the end user? The company has built separate user-interfaces for both HCP media buyers and media suppliers, including modules to specify lists of target HCPs, channels to reach the target HCPs (desktop, mobile, email, EHR, et cetera), timing, budgets, and bid/ask prices as well as how users would like to buy/sell, whether in a reserve (fixed price and volume) or spot market (variable price and volume) and various combinations with advanced packages with proprietary data to inform targeting and bid/ask price determination. The tool also offers several forecasting modules to give buyers guidance on what they can buy and when, and similar predictive analytic modules for sellers to ensure that they offer their inventory in a optimal manner to specific buyers.

How does it work? “The key components of the platform are (1) a multi-parametric matching engine, (2) a price clearing and yield optimization engine for both reserve and spot transactions and various combinations that market participants configure of which there can be tens of thousands of combinations, and (3) a transactional executional and settlement engine,” Gilbert says. “At any given moment in time, there can be more than 20 million transactional scenarios that can occur with multiple buyers, suppliers, units, and unit variables, and the key is to execute all of this within 30 milliseconds, which is faster than a person can blink their eye. Failure to transact within this high-speed timeframe can lead to meaningful losses in revenue and, as a result, speed and accuracy are paramount.”

According to Proclivity executives, the advantage that LayerRx offers over its competitors is how much “smarter” it is, and how much more information is at the fingertips of the user. “Most of the buying and selling activity that occur in marketplaces, especially online ad exchanges, is often based on largely arbitrary and subjective decisioning which determines what is bought/sold, when, how, and at what bid/ask price,” Gilbert says. “Given this, we have instead focused on giving buyers and sellers the appropriate modules along with new levels of transparency and data to determine, for example, the optimal circumstances to trade in a reserve or spot market as well as the optimal counter-parties to trade with for a desired transaction. Moreover, we are also allowing entities to appropriately leverage first and third-party behavioral and transactional data to effectively compute the value of an HCP to maximize their bid/ask valuation computations instead of manual and arbitrary methods to drive their digital programs.”

And there’s more “smart” on the way; Proclivity’s leaders are hoping to make the tool even more intelligent and predictive. “We have been excited to see, in a relatively short period of time, how many different digital channels have been added to the marketplace and think that we are only in the first-inning with regards to the type of ‘intelligent agents’ that we can provide to buyers and sellers to allow them to discover and activate the most effective transactional packages that will maximize yield to both sides using data and machine learning,” Gilbert told Med Ad News. “One of our publishers got a glimpse of where we are headed and when he connected the dots, he said, ‘Wait, you’re saying LayerRx can act as my Alexa and eventually suggest the best ways for me to sell, what to sell, and to whom?’ While we can not predict what lies ahead, we certainly think there is something there worth exploring as it relates to the future of online marketplaces.”