11 February 2020
The advent of data analytics in pharma marketing automation
In marketing to physicians, the days of “spray and pray” are, if not totally behind us, then definitely numbered. The spray and pray approach saw marketers complete their market research, create the message they wanted to send, and then fire out an email to everyone on their list, with no attempt to personalize the messages. The uptake of artificial intelligence (AI), however, in targeting and personalizing messages has been changing this practice. Now, when marketers send out an email, says Pratap Khedkar, managing principal, ZS Associates, “they can actually tell who opened it, when they opened it, whether they clicked through to the second page or click on any of the links.” Accordingly, the ability to use these data and insights to understand what customers need, what kinds of information they want, and how they want to consume is driving a marked improvement in the way pharma engages with physicians.
For James Anderson, chief customer officer at Aktana, a company providing AI-enabled decision support for the global life sciences industry, “the growth in this area has been insane.” Aktana’s biggest challenge, he says, “is keeping up with the increase in customers we have.” While the use of AI in marketing is not a new concept, what has changed over the last couple of years, explains Paul Shawah, senior VP of commercial strategy at Veeva, is the ability “to bring data together from multiple different sources and in a way that’s ready for AI to generate insights and guidance on messaging for specific customers.” Anderson agrees: “Ten years or so ago, pharma companies were starting to get their data warehouses together. Now the Pratap Khedkarlarge companies, especially, have a lot of data that is ready to use. Combine that with the advances in AI, analytics, our ability to compute things quickly, and that leads to the big rush we’re seeing now.”
For Khedkar, harmonizing the way the marketing channels are managed has become a crucial factor. “The job of sending these things out channel by channel used to be outsourced. You had an email vendor—all the emails got sent to them along with the target list and they managed it.” But with another party dealing with, say, iPhone alerts and another dealing with the reps, there were two or three other channels being managed independently.
“Companies have realized that you have to harmonize, which means one entity has to be in control of all these different channels to the same doctor,” says Khedkar. This is where AI is beginning to help, he adds. “The technology has given pharma back the control of how it can actually modulate when different channels get used, and in what order they get used with the same doctor. You cannot harmonize manually, because you don’t know what type of cadence, what type of timing, what type of sequences actually work with this particular HCP—that’s all data driven.”
With healthcare practitioners increasingly composed of millennials—for whom digital technology has long been not just a tool of work but, in many respects, a way of life—it should perhaps follow that the new developments in messaging are met with more active engagement by the younger generation than the baby boomers of old.
Khedkar has found, however, that “it’s much more about how many years the physician has been in practice, rather than the physician’s age.” Anderson adds, “I know physicians who are older and more tech savvy than me; I know younger physicians who aren’t as tech savvy. It all comes back to the individual preference. And we’re getting to the point where you don’t have to use proxies like age to figure those things out.”
A more digitally savvy audience is by no means an “easier” one, however. According to Krishna Kadiyala, vice president, head of commercial operations and innovation at therapeutic antibody company MorphoSys, “customers have raised expectations in terms of how they consume and how they disseminate knowledge.” The explosion of digital and social media channels, he says, is forcing companies to redefine their engagement strategies and shaping how they can be better partners to their customers.
Khedkar has observed variances in specialty more than in the ages of healthcare professionals (HCPs). “For example, rheumatologists and dermatologists remain very accessible to sales reps,” he says. “On the flip side, oncologists are much less accessible. They are much more likely to open emails and much less inclined to talk to sales reps.” Khedkar notes that dermatologists are still “very open to talking to sales reps, but their email open rate is only around 4.9%.” He adds: “AI and data has to be able to exploit all this information.”
Sales reps may be relieved to hear that certain specialty physicians still very much value their physical presence. Indeed, in the face of increasingly sophisticated automated marketing, reports of the sales rep’s death have been greatly exaggerated. Large pharma is getting more comfortable with the idea that this technology is helping to make the rep more effective, says Khedkar.
While sales rep numbers have dropped compared with a decade ago, he doesn’t see any further major disruption to the numbers. “The access situation has stabilized, and the number of reps now is about 70,000,” explains Khedkar. “The peak used to be about 103,000, so it has come down a lot. I don’t think the numbers will continue to drop—the 60,000–70,000 is here to stay.” But with AI’s mark on the sales and medical field teams proving to be an indelible one, the rep role will have to evolve. “Not by leaps and bounds,” Khedkar goes on. “Maybe 20% to 30% of the reps will not be able to do their job all that well. But I think the majority of them will have to adopt this technology so they can use it as an additional arrow in their quiver.”
Indeed, at the heart of the pharma–physician relationship, the rep’s role will remain vital. Khedkar points to a statistic about emails sent to doctors. If the email is forwarded to a doctor by a rep—with the rep’s email address and signature—instead of a third party, the doctor is six times more likely to open it.
For Shawah, AI is having two big, positive effects on the sales role. One is the “reinforcing of goodPaul Shawah behaviors.” He explains, “The AI initiatives that have sprung up over the last four to five years have generally coincided with field teams having access to more digital ways of engaging with their customers, like remote meetings. Before that, most field personnel did not engage with their customers digitally. AI has reinforced the use of these new channels and when to use them, how to use them, and how they become part of their customer engagement process.”
The second impact is where AI surfaces insights that the customer account team “would never have known,” says Shawah. As an example, “a key account manager may not be aware of something that has happened with their customer, or even the patient; perhaps they get an alert that patients associated with a particular specialty pharmacy are delayed receiving their medications because of a reimbursement issue. Using AI to surface those insights, the account manager can then intervene and becomes more of a trusted advisor.”
Salvatore Paolozza, director of sales operations at Antares Pharma, agrees. “Basically, the intelligence that we’re giving to our reps is a lot further advanced than it was before,” he says. Antares is an adopter of Andi, Veeva’s AI application for customer relationship management. “What the application brings,” Paolozza explains, “is greater flexibility and greater interoperability between sources. It completes the picture, if you like.” Crucial here is AI’s ability to learn, he says. “We’re moving out of a coding environment to more of a learning environment. These applications can learn with time what they should and shouldn’t surface up.”
To stay in tune with the advances and possibilities of new applications and algorithms, pharma is keeping a close eye on the activities of the big tech companies.
No one can attend a conference on using technology to communicate with HCPs without hearing about Netflix, Amazon, Apple, Google, or, to a lesser extent, Facebook. Pharma is some way behind these digitally native companies, of course; says Khedkar, “that’s partly why the industry is looking to them. The question is, what else can it adopt?”
Some of the algorithms coming into pharma marketing and medical communications are indeed being adopted from Netflix and the like. Khedkar points to an algorithm called collaborative filtering, which is the way Netflix figures out what movie to recommend to individual viewers. “We took that algorithm, which is in the public domain, and modified it to work for pharma to identify specifically what type of content to send to a particular physician,” he says. “As well as figuring out which channel you like, it’s crucial to figure out what content to send. Should you send an efficacy message? An affordability message? A safety message?”
The problem with adopting such an approach, however, is that pharma “has gotten spoiled with sales data; they think that if they don’t have 90% of the data right, then nothing can be done,” says Khedkar. This is a mindset that needs to change, the executive believes. “What data do you think Netflix has?” he says. “When Netflix looks at the ratings from people who watched a certain movie, they have 2% of the data. What they’re doing is extrapolating the 98% they don’t know from the 2% they do know.”
Pharma has to get more comfortable working with “poorer or incomplete data,” says Khedkar, and let AI do its work. “The AI team is important because the data is not going to be perfect. Pharma needs to be a little bit more courageous in using AI on the data it has, because AI will compensate for the holes in the data.”
Of course, pharma marketers are dealing with a considerably smaller audience than Netflix, maybe a pool of 10,000 physicians rather than millions of viewers. But for Khedkar, the “spray and pray” approach that turns physicians off is a much bigger risk in the smaller pool. “Both extremes are bad—hesitating and not doing anything because it’s a limited pool or opting for spray and pray,” he says. “You have to avoid both extremes and find the happy medium, which is where AI and data comes in.”
While attempting to emulate some of the Netflix or Amazon methods is ambitious, pharma can look to other, more regulated industries for a more cautious approach. Says Shawah, “We often look to financial services because they are one of the best industries at identifying really valuable, high-net-worth customers and then managing them through the relationship life cycle.”
He adds that in the telecom industry, the wireless service providers, for example, have a wealth of information about a customer before that customer walks into a physical store. “They know everything the customer’s been doing on their phone; they may know if the customer is at risk of moving to a different carrier, or whether the customer is over or under his or her data limits,” says Shawah. “They’re able to use that information to have very targeted conversations.”
Some industries are not always relevant, adds Shawah, “but it’s important to look outside of life sciences as much as we like to look inside of the life sciences,” he points out. “There’s no monopoly on good ideas and where they come from.”
All the advancing algorithms and technologies notwithstanding, the pharma industry must remain mindful of “the human being in the middle.” In engaging HCPs, drugmakers are not dealing with a machine-to-machine interface; it is machine to human. “Because of this linkage, the data, the interactions, and the nuances will never be perfectly captured,” contends Khedkar. AI provides “just a different model.”
There will always be an element of complexity and imperfection, he adds. “It’s about balancing the personal and the non-personal.”
The key to success in using AI to engage physicians is making that marriage more fruitful.Print
28 September 2020
28 September 2020
24 September 2020
24 September 2020