Leadership Exchange: Social Listening
Marc Iskowitz (editor in chief, MM&M, left): We've got a great group here to discuss the opportunities and pitfalls in leveraging social to become more customer-centric. Let's start by having each person give his or her assessment of where they see the pharma industry along the social-media adoption curve and what catalysts they think are propelling greater use of the medium.
Tim Armand (co-founder and president, Health Union, right): We create communities dedicated to specific health conditions. I used to joke even five years ago that social media was the unstoppable force and pharma was the immovable object and what's going to happen when these two collide? What's driving people to do it is the … rising tide of patients there. People are on social media, they're engaging and … there's a lot of downside of not listening when this conversation is happening. So doing nothing is not the right answer. There are just way too many people participating and talking about their health, specific conditions and specific treatments on social media. What's also driving it is it's no longer a niche. It's no longer just for the kids. The demographics on Facebook would support that.
Siva Nadarajah (general manager, big data and compliance, IMS Health, left): I head up the social business, which also has the big-data business at IMS right now. My company [Semantly], a technology-services company for social listening and compliance, was acquired by IMS in 2013. When we started Semantly, in 2010, we … approached it from “we need a vertical solution for life sciences with compliance embedded in” so that we could go and talk to the regulators, go and talk to the pharmacovigilance [PV] folks first, convince them. Then you go to the brands and others and say, “This is being looked at by your PV. They're fairly okay with it, now you can go do it.” We're now able to use the platform to help take social media to an actionable state. Social listening has always been something that was done on the side. Now … it's very, very important. It's not something you can ignore, not a vocal minority.
Wayne Pines (public affairs/regulatory, APCO, right): I'm a regulatory consultant and I was formerly associate commissioner for public affairs at the FDA. We already have seen over the past few years, and we're going to see with greater intensity over the next several years, many more opportunities for patients to participate in the drug-development and drug-approval processes, product registries, natural history studies and patient-reported outcomes. How social media plays into that remains to be seen. What the FDA always wants is data, not anecdotes. So how can we use social media as a vehicle for collecting appropriate data that can actually advance the science of understanding a particular drug or a particular disease? A number of companies have recognized that trend and are -bringing in patient representatives to various meetings to talk about the disease and what patients' expectations are in terms of treatment.
Michele Bennett (senior director, data science and analysis, Thomson Reuters, left): Prior to joining Thompson Reuters, I headed up research for a company called Wool Labs. We were a pioneer in 2007 for creating a [pharma] social-listening platform and analysis system. Pharma is on the very far right-hand side of the adoption curve. Why is this growing? We have found over the years that sometimes social media is a predictor, sometimes it corroborates stories, sometimes it's lagging behind, but it's always part of every launch, every product, every disease state … even the smallest disease state. It's now an embedded part of the health landscape. Pharma also is used to being in control, driving the message, driving how they want their products to be perceived, within label. Social media takes the message over. You get no control.
Greg Cohen (associate director, global strategic marketing [multichannel], UCB, right): I've been at UCB for a little over three years. I worked in US social media for the bulk of that time and then about five months ago switched over to a global role. From a pharma standpoint on the adoption curve, there's so much opportunity for pharma, which is such a scientific and technical industry, to really make things so much simpler to consume for patients and physicians in these mediums, whether it's through video, infographics, Facebook or other online content, as well as via six-second videos, 15-second videos—whatever bite-size content is popular. It really is something that plays into the strengths of what pharma can communicate. Unlike something like a CPG or retail, where how much do you really need to know about the shoes? But in this case, having this opportunity to now tell these mini stories around content is something that companies should be embracing.
Marc Iskowitz: Now that we've discussed the catalysts, what percentage of brand managers are really in patient-engagement mode with social, or even listening?
Siva Nadarajah: Last year we did some research on how many companies are engaging with patients or at least thinking about engaging. It's around eight to nine percent of brands—not companies—globally. The eight percent is concentrated mostly in the US and also includes Brazil, Turkey and South Korea, which are some of the largest social-media countries where companies freely engage. China is starting now. Communities are different there. There's no Facebook.
Marc Iskowitz: And the other 92 percent?
Siva Nadarajah: Some are at least listening and doing nothing about it because they don't have a way to communicate or engage. About 50 percent of them are at least listening. Another 40 percent are still not even listening. It's much more in Europe, especially. The percentage of people listening there is very low. Japan is a place where they don't listen because there's no content there. So there are countries where … you can't listen.
Marc Iskowitz: What are the reasons why the other 40 percent are not even listening?
Siva Nadarajah: Primarily, I would say the brand managers who have the data but still are not listening are the ones who [fall] back on regulations and say, “I don't want to deal with adverse events. It's a burden.” The others say they don't know what actions they could take with the data. Then there are others who know what to do with this data but they've been tied up because they don't have engagement channels to take the next action—they know what to do but can't do it. But the eight percent of people who are engaging are definitely listening, because without listening they can't engage.
Marc Iskowitz: Is compliance the barrier to social participation that it used to be, Wayne?
Wayne Pines: I understand that companies are having a tough time trying to comply. That's what I've heard so far and that is the main barrier to expanded use of social media. . . . In terms of the general environment for compliance, most of the major companies are under corporate-integrity agreements, which means that they have to be absolutely sure that they're compliant. Not that they would not try to be anyway, but they'd have to certify it. I think that's brought a lot of pressure on the companies to approach any new media. There's a view that the FDA has not yet set forth in exquisite detail exactly how to comply and what the parameters are for social media. [The last time that the FDA issued draft guidance for social media was its June 2014 proposals focused on presenting risk information on Web platforms with character limits, and another dealing with misinformation about products by third parties. —Ed.] The understanding by the companies of what is permitted and what is not permitted will continue to evolve as it has with DTC advertising in the past. If you go back to the beginnings of DTC advertising on television, which started in the early '80s and it wasn't until 20 years later that we had a very, very clear understanding of what was permitted and what was not. So I see it as an evolutionary process.
Marc Iskowitz: Corporate understanding takes time to evolve. Do we all agree that this is a barrier?
Greg Cohen: One of the issues companies run up against is they don't think they know what they're allowed to do, which I usually push back and say, “We know the rules.” The FDA kind of put out “here's some guidance and in lieu of no guidance, some guidance is better than nothing.” Then what I tell everybody after that and I think is sort of my mantra now moving forward is sort of the original Google mantra, which was “do no harm.” This idea that if you're doing things for the right reasons with the right intention—we're not trying to deceive or sell you something you don't need. If you're doing things for the right reasons and the right ways and, Wayne, you can tell me if I'm wrong, but that's in line with what the FDA wants. They want you to be fairly communicating the benefits and the problems and they want you to present something in a way that patients are actually consuming.
Marc Iskowitz: What's UCB's process for adverse event [AE] reporting?
Greg Cohen: We've had a very progressive digital PV group that has been on board from the very beginning in talking about why are we doing this, setting up the right parameters to handle the inflow and automating what we can. Our work with IMS has been extremely helpful in processing and filtering out and avoiding duplication of reporting. So we've had a very good team that was able to take a step forward and say, “Let's take a proactive stance on making this work.” That has made a big world of difference.
Tim Armand: How do you handle anonymous or semi-anonymous AEs?
Greg Cohen: If we have an avatar or any sort of name that we can tie it to, then we consider it a name. So the four things you need are the name, the drug, who's reporting it and what the AE was. If we know it's “GCohen85” but we don't know anything else other than a screen name, we consider that a name and report it as the name. Then if we ever in our future have a way to triangulate it, we'll make that connection. They say, “I'm on this drug and this is the situation I had.” We take a proactive response in terms of reporting all of those. If we get more information, we can, but obviously if it's in a forum where we have no presence, we can't follow up with them. If they say it in one of our communities, we post information for them to be able to go and do actual reporting for it. Now that we're in the Twitter space, if people have issues we respond to them and say, “It sounds like you're having an issue with one of our products.” We record it on our side that this person tweeted out this issue and then we put it into our system. So if they ever contact us back through the system—they say, “You tweeted at me about this, here's my handle”—we can go fill in their information and do the whole intake process. It's not perfect yet, but it's a first step and we're getting better at it.
Marc Iskowitz: Are there areas where you'd appreciate more clarity?
Greg Cohen: My only issue with some of the more recent FDA guidance is around viewing these mediums in the older way as opposed to how patients and online consumers are actually using them. For instance, some of the restrictions on Twitter are really hard because with 140 characters you can't do anything. So people are experimenting now with using images that have the safety text, and there are a lot of workarounds, but that's also not the way that patients are using the medium. So it's not necessarily doing them a service.
I think the people who complain about it don't understand the box you have to work within. Everybody loves to think outside the box, but pharma is very much about how do you work within the box, not in a limiting way but in a way that makes physicians, patients and regulators comfortable. That's where the real challenge and opportunity are to me. What is that box?
Marc Iskowitz: Where else could resistance be coming from?
Tim Armand: It's the fear of the unknown as it relates to regulatory and legal compliance. It's also—again, I'm a 13-year veteran of pharma so I've been there, done that—but doing anything new is hard because you have to, by definition, display something that was a great idea last year because of the zero-sum budgets in most cases. So there's fear of displacing something that was working well in the past. There is the fear of the so-what factor—can I actually do anything with all this wonderful insight?
Greg Cohen: Another obstacle is that social-listening outcomes are inherently unplanned, and the output might lead to more unplanned expenditures. We don't always know what it's going to be, what it's going to look like, what we might be doing. I could put in the budget to do social listening, but where does it go—is it an R&D thing? Well, R&D doesn't have money for that. Is it a commercial thing? Is it online advertising? It's not that we don't have the money for it, it's that this is an unplanned expense. And who knows the size—bigger than a bread box, smaller than the Empire State Building. And the other question is, how do we know this is a real issue vs. a spike? Maybe we should wait to validate it in six months and see if it's an issue. The unplanned nature of what you might learn and what you might be able to do with it throws off the way most companies budget for activities. Pharma's so rigid in how we plan stuff.
Michele Bennett: As well, there isn't consistent ownership of this, so it doesn't have its own budget. It doesn't have a history of saying that “every year we do insights mining,” and then we know we have a road map. If it was institutionalized, it would have a budget. So it's going to struggle all of the time because then you're trying to steal budget from somebody else who thinks their stuff is important, and rightfully so.
Marc Iskowitz: How do traditional marketing research and other groups feel about this relatively new form of insight, and how does it complement other kinds of analysis?
Michele Bennett: We treated it like a research project, and that took people off guard at first. The market-research folks were like, “You can't treat this as a research project.” I said, “Let me give you 25 articles from peer-reviewed journals that treat it as a research project in academia. So we can do that, too.” It changed the whole tone. It's also faster and less expensive [than other forms of analysis]. So if you did a fast surveillance study and then that could inform your panel groups, why would you not want to know some things in advance? It can confirm what you learned with 12 people over a larger population. They're complementary and different at the same time.
Marc Iskowitz: What's the ideal “actionable state” for brand teams to reach with action-oriented listening?
Siva Nadarajah: I normally ask clients, “What are you going to do with this data?” If they say, “I don't know,” then I would suggest they rethink the whole thing because we don't want to be in a situation after three months that we produce some data and you sit with it. It's a project for us, but it's not an optimal relationship. So the ideal state is to know the endgame, have an action plan and engage, because the guys who do the listening may not be the ones who could take the action, which would be the brand marketers and the folks who can manage the websites. A lot of things can be changed on the website based on listening. We find all the time folks, patients especially, looking for information that's not available on the website and then they go to Wikipedia.
Tim Armand: The mitigating approach to that is a sort of a laddering approach. Try something small—quick wins. Get people comfortable with it, partner with your internal colleagues—legal, regulatory compliance, etc.—early on, not after the fact. To Siva's point, have a goal in mind with what are you going to do with it. When I was running a brand team at GlaxoSmithKline, we'd sit down with the market research people and line item by line item go, what are we going to start, stop, or do differently because of this research. If the answer was “we did it last year,” then it was … “what is going to be the change?” Agree up front.
Greg Cohen: This is all about laddering wins. So obviously you want to get the quick ones you can get done. We try to find the stakeholders in the company who are looking for answers to quick questions. So we've done a couple projects, two-, three-week little sprints of like, “I wonder if people are talking about this in social media?” We would do a two-week run of looking at the last six months or so of data and come back with, “Oh yes, we found 400 conversations out of our larger collection that talk about this.” Then as we dive deeper and deeper into that topic, you can kind of see some different topics coming up. So we can very quickly come back and say, “Here's the matrix of conversation that we're seeing.”
Marc Iskowitz: At the end of the day, you're drawing conclusions from unstructured data. Where are some of the socially valid uses of this information?
Greg Cohen: What people are saying online is also as informative as what people are searching for online. You can learn from the balance of the two. For one of the new indications that we had launched a couple years ago, we could see that people were searching for images of what a disease looked like in the hands. So in our materials, we started including pictures of hands. The next piece out is, who is the right audience? Are we even talking to the right people? We can learn so much from how they talk about things that we can then [use to] segment. This is like a level of secondary segmentation. It wouldn't replace all the work you do up front in determining who's the right patient for this product based on clinical trial information. This is in the real world. At the next level, I put R&D, the idea of taking what you've learned and asking how we can make the product better for these audiences. The final layer would be taking that information and moving it into whole new markets, getting into the molecule-investigation process and determining if we are missing people who have a totally unmet need.
Siva Nadarajah: The other group of thinkers that we are seeing in the space is the real world evidence group, the HE-OR type of folks, who are using this data, then taking it to the next level. Social is going to give them some signals on switching behavior. Then to validate that, they run surveys either in their communities or through other channels, then take this data and anonymize it and go back to the patient-level data they have.
Marc Iskowitz: Staffing and level of internal social-media expertise are two other reasons why the industry may lag. Greg, who should drive the discussion internally?
Greg Cohen: It should start with a holistic customer-experience marketer.