Episode 43: Driving Real Results Through Data Privacy and AI in Healthcare Marketing

In this episode of the Digital Clinic, Todd Weinstein, Director of Digital Marketing at NeuroPace, dives into the real mechanics behind driving results in healthcare marketing. Todd shares his frameworks for connecting digital campaigns to business outcomes, how he uses attribution models across patient and HCP audiences, and why the discipline of proving marketing’s ROI has never mattered more in a compliance-driven healthcare environment.

What makes Todd’s perspective particularly valuable is the combination of depth and range he brings. He has lived the full arc of digital marketing in highly regulated industries, from agency life across pharma and MedTech to his current role on the client side at NeuroPace, a medical device company whose RNS System helps patients with drug-resistant epilepsy reduce or eliminate seizures.

Todd also shares a perspective that we at Wheelhouse have said time and time again: the privacy constraints healthcare has wrestled with for years are coming for every industry, and the playbook healthcare has been forced to build may be the most valuable roadmap in digital marketing right now.

Listen & Subscribe:

Career Journey: From Agency to Client Side in Healthcare Marketing

Aaron Burnett: You have had such a varied career and an interesting career. You really have lived the evolution of digital marketing in highly regulated industries in pharma, then biotech, now MedTech. Can you just give us kind of an overview of your career history and the different types of industries and companies you’ve touched along the way? 

Todd Weinstein: Sure. So it’s certainly been, like you said, all within healthcare started out on the services or the agency side, and so they’re really introduced to a myriad of therapeutic classes as well as ways in which to market drugs and devices to patients, caregivers and physicians, and then any other kind of tertiary stakeholders. 

The industry has certainly evolved and grown up over the years, and then the past several years I’ve joined with what is called the “client side,” right? So sitting in the seat of a digital marketer making some decisions in terms of how you reach patients and caregivers through digital channels as well as HCPs. 

And then through that evolution have tied things more closely to sales in terms of the output of the digital marketing activities. A little bit more so than I had paid attention to when I was on the agency side. So those were the biggest differences in terms of at least my own personal journey from a digital marketing perspective and then obviously all the tech and evolution of digital and healthcare has exploded.  

Tying Marketing Activities to Business Outcomes

Aaron Burnett: You’re speaking my language in terms of time, marketing activities to sales outcomes. Most people would tell you that’s very challenging, particularly after the OCR guidance in November of 2022. 

But it sounds you have been consistent, I think, in your ability to do that. So let’s start there. How are you able to tie the efficacy, the value of marketing activities to true business outcomes?  

Todd Weinstein: Yeah, so you’re right. It is difficult, and I won’t say it’s a rather inexact science, but in order to quite honestly justify your marketing budget and continue to do the things you want to do, you’ve got to somehow tie it. 

Either to a bottom line revenue figure and or certain lower funnel activities that a patient or a HCP would take on a website or social media, whatever the conversion medium is. I’ll start on the patient side because that’s more of an inexact science than it is on the HCP side, because obviously patients cannot write their own scripts. 

They’ve got to go in and either ask for or about product. With patient, it’s really tying things to what we would consider highest value action. That could be getting them to sign up for some sort of CRM email marketing campaign where you’re now able to get name, email, and potentially address to be able to have a longer tail communication stream with them over several months, years, however long that journey may be to provide to them either information on drug efficacy to get them to get on drug or remain on drug, or maybe it’s adherence or other. 

Events that you want them to know about the product. You’re also driving through specific lead gen channels, getting them to potentially download specific assets. The typical one is a download of a doctor discussion guide, because that’s the closest thing you can get to a buying signal that they’re interested in the drug. 

They’ve downloaded the guides that they can ask their doctor about it. So usually using any kind of paid media and attribution to understand how many patients are doing these things that you’re paying for, and then determining your ROI from there and determining some sort of lifetime value for the patient if they engage with these activities. 

On the HCP side, it’s way more exact, right? They’ve got NPI numbers aligned with them. You can get claims data that tells you how much or little they’re prescribing your drug or device in the category. The only caveat I would say is that you still don’t know beyond the shadow of a doubt that it, it was marketing piece A, B, or C that drove them to a decision. 

And then obviously you’ve also got your field out there calling on these physicians and they’re certainly influencing treatment decision and pulling sales through. So that part’s easier because we know which doctors are writing what. But in terms of where to attribute the revenue, it’s a little bit more challenging because you’ve got to create business rules that typically split it between, I’ll just use a hypothetical, 50% value for the advertising you deployed and 50% for the field. 

Aaron Burnett: So you’re going for an implied attribution model?  

Todd Weinstein: Correct.  

NeuroPace and the RNS System

Aaron Burnett: Can you tell me about your role with NeuroPace and tell me a bit about NeuroPace?  

Todd Weinstein: I’m the director of digital marketing at NeuroPace. In short, that kind of oversees all things digital performance or measurement performance in terms of how our campaigns are doing, the website, all kinds of paid media, including social email marketing to both patients and physicians. 

Really any kind of non-personal promotion that goes out to the patient, which is everything because we’re not calling on patients. And then also helping to support the field with any kind of digital air cover that would help them in their efforts to be able to get access to physicians. 

And then obviously talk to them about our device, which is the RNS system. NeuroPace is a technology forward organization just based on the device itself. The RNS system is a device that recognizes and either reduces seizures and or heads them off at the pass. So it’s a device that’s implanted through a procedure. 

Also has the ability through technology to be able to provide. Real time data back to the physician in terms of efficacy and how it’s doing per patient to understand if any adjustments needs to be made, everything is programmed the way it needs to be working within the neurology field. Obviously amongst neurosurgeons and epileptologists are the call points, and then certainly serving patients to help them as best we can to be seizure free. 

Aaron Burnett: It seems like an amazing device. I was reading that in order to qualify the patient must. Have not had success with two other epilepsy medications. And so I would imagine that they come to you and this is glass gas, desperate for a solution.  

Todd Weinstein: So yeah, drug resistant epilepsy is the nomenclature. 

So yes, they have to fail at least two medications prior to being eligible for the RNS system. So when they arrive at the epileptologist, in terms of cycling on and off specific medications, this is an option for them that based on our efficacy, works quite well. There’s different levels of that and certainly different success stories that they achieve based on the RNS system. But our hope is that obviously it, it helps them as best we can.  

Building the Right MarTech Stack

Aaron Burnett: Given your focus on measurement and proving out the value of marketing, talk with me a little bit about your approach to your marketing tech stack and analytics. You are quite often the first marketing hire, so you are coming in and setting the stage and probably making infrastructure decisions. What sort of tech stack do you use now? What is your ideal tech stack?  

Todd Weinstein: Yeah. You could go a lot of different directions, that’s for sure. Some of it is core block and tackle standard, right? Google Analytics. In terms of the website metrics, we also work within Salesforce and then Tableau. 

From a data visualization perspective, I think the challenge that we have in probably a lot of organizations is that typically the data is coming in from disparate sources, and so you’ve got to be able to aggregate it into a uniform look and feel to be able to report out on it, and then obviously make changes based on what the metrics are telling you. 

It, as presently constituted it’s a lot of the usual suspects in terms of CRM as well as automation, typically an omnichannel, Salesforce is the, at least in healthcare, omnichannel Salesforce is usually the choice. Given that it’s probably being used in other places within the organization, and then it can seamlessly integrate with other tools. 

HubSpot is another one that can also be used from both a measurement and an automation perspective. You’ve got to have all that data flowing into a central place, including the digital media performance. So that you can aggregate it as best you can and not report out on it separately because the reality is that your digital media is driving your website performance and landing page conversions and optimizations and or your website needs to perform in a certain way. And that could be impacted by landing page specifics, which are measured in Google Analytics. So it’s all got to integrate and talk to each other to make it work.  

Aaron Burnett: Because you are in a highly regulated environment. Have you needed to change the implementation, customize the implementation of GA4 in particular? 

Todd Weinstein: No, we haven’t had to change anything in Google Analytics. No one’s really coming through with identifiable information. It’s more patients get tokenized in other settings where there’s a patient ID associated with the patient, so that. There’s no personally identifiable information that is divulged, and almost every, if not every healthcare vendor or partner has complied with that rule for quite some time now, because there are regulations obviously on the pharma device and biotech side that call for those things. 

Aaron Burnett: You have frequently been the first. Hire for digital marketing. What do you like about being the first hire? What’s exhilarating about that?  

Todd Weinstein: Yeah, I think it’s something that obviously you get to build not always from the ground up, because there’s typically some sort of digital infrastructure in place because you can’t market really anything, whether it’s this industry outside of it, without some sort of digital footprint. 

But it is the chance to build something, put a structure around it, put some objectives and some goals, both short and long term in place. That allows folks to understand what the path is and what we’re trying to accomplish, and then how we’re going to measure it. And then I think it also goes back to what you said in terms of tech stack. 

Typically you’re able to implement a tech stack that you feel is best for the here and now, but also has the opportunity to grow, whether it’s expanded licenses or footprints or digital channels that allows you to operate in specific places that maybe you haven’t been thus far. I think it’s really the ability to set forth a vision. 

Kind of own the path towards that vision and be held accountable. Because if you’re building something for the first time, yes, it’s exciting, it’s exhilarating, but there’s also a certain amount of expectations and pressure to be able to deliver.  

Aaron Burnett: You’ve been agency side, you’ve been client side. 

Todd Weinstein: Correct.  

Aaron Burnett: What did you learn when you came client side that maybe was enlightening, was a surprise to you versus the perspective you had on the agency side?  

Todd Weinstein: This will sound a bit counterintuitive as a digital marketer, but on the agency side, there was this obsession with innovation and trying new things, which you do want, right? 

You don’t want to just copy and paste your marketing plan from year one to year two. There has to be some sort of change, whether it’s different channels, different investment levels, brand objectives, et cetera. What I found on the agency side was that obsession with innovation and change wasn’t always embraced. 

On the client side where I am now, and part of the reason it’s not always embraced is that unless you can tie that innovation or that change to sales and bottom line, it’s typically not going to get funding and backing. Part of me understands that for sure. It’s a business and so if you can’t prove out innovation in terms of how it’s going to net new sales or at least project or predict it, then you’re not going to invest in it. 

But the other side of me says that’s what innovation and digital transformation is. You’ve got to try and do some new things based on the way the industry potentially is moving, and that’s how you set up a test and learn. On the agency side, it was more innovation and change and trying to sell that in, whereas on the client side or the manufacturer side it’s not always viewed that way. 

When you’re on the marketing side, there’s a very keen and consistent eye on sales, especially if you’re working as I am now and in the past. For a company that has one product or one device, you’re living and dying by the sales of that, your marketing activities, your budget, and just your day in, day out for the most part. 

Is keenly tied to that sales figure, whereas on the agency side, you’ve got multiple clients. Some client sales could be good, some could be bad. There could be market dynamics, and it could be all over the place so you’re not as in tune with that sales number, even though it does impact your business as an agency, you’re not sinking and swimming because you do have other clients that you typically have a book of business.

Attribution Modeling and Long Sales Cycles

Aaron Burnett: With the prevalence of cookie deprecation, increased privacy regulations and suppression, and the loss of signal that comes from that. How are you approaching attribution modeling, if you’re using it. Are you shifting to something else like Media Mix Modeling to get a better understanding of efficacy and then make decisions with regard to how to place media bets?  

Todd Weinstein: Everything you mentioned is all on the table, some of which I’ve done and I’ll get into more detail. 

So in terms of the third-party cookie deprecation, really what you want to try and do is establish first-party data relationships or first-party relationships with your physicians and patients. What do I mean by that? You can still use digital media as a lead gen tool to get people to, like I said earlier, sign up for CRM or sign up for email. 

If you can do that, now you’ve got first-party relationship with the patient or the physician to be able to send them whatever you choose, right? Whether it’s efficacy, safety, adherence information, whatever it might be, a new drug launch or extension. So that has been a shift that’s been in place for some time to be able to try and get that first-party data relationship or first-party relationship. 

With key customers and constituents in terms of attribution modeling and how and where to attribute revenue based on marketing activities. It really comes from business rules, right? So I’ve worked on brands before where we’ll run a campaign and we’ll say, okay, any script that comes in 30, 60, or 90 days after the physician having engaged with the campaign. 

We’re going to attribute a certain percentage of the revenue to that campaign, whatever it might be. But after that, we draw the line in the sand and say it could be something other than this. Because they’ve had, a fair amount of time to show a behavior change in writing, and now that window has elapsed. 

That’s variable based on the patient journey. Some products you’re diagnosed and written a prescription for whatever drug it is, sight on scene that day at NeuroPace, the patient has to fail to epilepsy meds before they’re eligible for something like the RNS system. So that takes some time. 

So it’s a little bit more difficult in my current setting to attribute a specific campaign directly to. A patient getting the RNS system or a physician writing, the RNS system, just because the journey is significantly longer. But at the end of the day, whether it’s patient or HCP, you’re creating business rules that every level typically the organization has to agree on in terms of the attribution that you would give to that campaign or that media tactic getting a sliver or all of the credit for the script.  

Aaron Burnett: We’ve dealt with similar issues with MedTech clients and the way that we solve for that is to develop lead scoring rules. We have a high volume of customers on leads who become new customers. We developed the lead scoring rubric so we understood the signals that indicated a high likelihood of conversion, and we then converted that into a propensity score that we implemented at the moment of lead capture and sent. 

We send back to the platform’s real time via third-party like Tealium or Segment, or you could even do it via GTM, so that we are optimizing for. Every lead, every time a lead comes in, that was good, that was bad, that was middle. And that made a profound difference in outcomes for us. And similar situations where the conversion cycle might be as long as a year, and so you needed that lead scoring mechanism at the moment of lead submission in order to optimize campaigns in real time.  

Todd Weinstein: Yeah, no, that makes total sense. And yes, I do like lean scoring as well, because. It’s a proactive way in which to then more smartly invest money down the road versus other measurement types, which are retrospective. 

It’s good to know what worked, what didn’t, so that you could optimize. But at that point that money is sunk and if something failed, you can’t go back and course correct. You mentioned earlier marketing mix analysis. I’ve been with companies that have done a marketing mix analysis on the HT P side, which is interesting because then you’re taking the fields and the sales team into consideration. 

You’re putting a certain value on their impact to be able to drive revenue, which I would say most organizations, they are the largest revenue driver, right? They’re out there selling the drug or the device. You’re looking at that in concert with other digital media types, the website, and then you’re putting in the cost for all of those things. 

So there’s the cost for the field force. Obviously they’re making a salary and there’s benefits, et cetera. The cost to create and maintain your website, the cost on digital media. So that to me is a very interesting way, albeit retrospective. To determine what levers you’re pulling as you go into the next year’s plan to figure out where the investment is or is not. 

Aaron Burnett: One of the big shifts that we’ve seen that’s made Media Mix Modeling so much more valuable for us is that the model build time and the refresh time is so much shorter. Historically we started working on medium X bottling, three years ago in a context where. In most instances, the models took months to build and then once you ran them might take months to refresh and run again. 

And they were myopically focused only on paid media and often only on digital media as well. And we’ve been able now to build and use models that take into account all channels. Organic and paid, traditional and digital. And once we built the model and achieve the fit that we need, we can run once every week or two and get new signal based on new spend, new conversion, that sort of thing. It’s been a game changer for us.  

Todd Weinstein: Yeah, it makes a lot of sense because you are absolutely right that it takes a while to set up. Part of it is because you’ve got to get all the inputs and that it’s just manual labor for the most part, right? You’ve got to say, okay, here’s all the channels that we’ve operated in the last year, and here’s all the costs associated with it. 

The analysis part also takes longer than what you’re describing. Once a week is great because then you can start to again, make. Changes on a more consistent basis, based on what’s been put into the modeling process?  

Aaron Burnett: Yeah. We work with some clients where seasonality or even weather. Can have a significant impact on how spend plays out, how effective it is, and so we’ve incorporated live weather data. If you’re looking at live weather data and data changes in a region of the US you want to be running frequently and making adjustments based on the way the environment has changed. So I’m impressed with what we’re able to do now versus the platforms and technology we had to grapple with two or three years ago. 

Todd Weinstein: I won’t mention the name of the brand. This was when I was on the agency side, but it was it was an allergy product. It was an RX, it wasn’t an OTC, and that absolutely was seasonal. This was many years ago, so there might’ve been mixed media analysis that could have been done, although that didn’t happen for this client. 

But you’re exactly right. Basically, the product would follow certain seasonality throughout the country. Deploy regional marketing to those areas because you knew that once the allergy season was hitting in that region of the country, you should really pump up your advertising there and then decrease in other areas where it wasn’t the allergy season. 

Simple stuff but still helps your money go a little bit further when you’re marketing to people that are probably more in need of the product than not.  

Aaron Burnett: So continuing with data analysis and MarTech, how have you overcome the distortion of needing to. Evaluate data in platform. So every media channel is going to give you their own reporting data, and they all will bias toward their own self-interest and tell you that their conversions are the best conversions and there are more of them and they cost less, and all those sorts of things. 

To grossly oversimplify that, we’ve approached that by creating our own HIPAA-compliant data warehouse and normalizing that data, do you have a similar data warehouse or an alternative mechanism that’ll gives you this holistic view and allows you to control for bias?  

Todd Weinstein: Yeah it’s a good point. I have done that at other organizations where I’ll work closely with commercial analytics to basically ingest that data and then normalize it for our purposes. 

So really what you want to do is be able to synthesize the digital media data and then translate that as best you can to a conversion, whatever that conversion is defined as. The ones I talked about earlier are rather direct and traceable, if you will. The CRM signup. The download of a doctor discussion guide the 30, 60, 90 day lookback on a physician in terms of whether they’ve written the drug or not, or device. 

After being exposed to the campaign, that’s typically how I would approach it. I will say that most media buys are aligned with reach and frequency, which is fine. You still need a certain level of reach and frequency to be able to achieve whatever the goal is in the middle of the funnel. That’s education, and then the lower half, which is conversion. 

Really, any media company that I’ll work with or media publisher, I will always say to them, “Hey, yes, I know we need those metrics up top. We have to be able to, and we could do this internally on the client side, we have to be able to tie those to some sort of, of key action fulfillment.” If and when we do that and we’re not seeing the conversion we want, I really don’t care how many impressions I’ve served. 

If they’re empty impressions, then we internally aren’t hitting our goal. We’ve got to change course, whether it’s media or whatever the campaign is, and everyone solely use their metrics as the end all be all in terms of how we’re performing.  

Aaron Burnett: So let’s talk privacy regulations. We touched on that as we were setting the stage at the beginning of the discussion. Given privacy regulations, given the increased privacy regulations, how have those constraints affected your marketing strategies and tactics? How have they limited or maybe surprisingly improved the value and the quality of what you’re doing? 

Privacy Regulations and Compliance in Healthcare Marketing

Todd Weinstein: Yeah so anything we’ve produced, whether it’s at NeuroPace or any medical device company, any pharma, biotech has to go through a regulatory approval process before it is in the public domain. So medical, legal, regulatory, all has to take a look at it and approve. And part of that yes, is privacy. 

It depends on the organization in terms of the regulation that they’ll put in place. I guess from the patient and the HCP side comes down to consent, whether that consent is gained or not. And then if it’s not, and we’re not even asking, do we have the ability to put together, push marketing to that physician or patient, but allow them the opportunity to opt out? 

The simplest one I can think of is any email you send, whether it’s the patient or HCP, whether it’s a push email and they haven’t signed up or they have. You’ve got to give at the footer, the option to unsubscribe. It’s the Canned Spam Act so that you’re not. Violating law, basically. Yeah. I’ve worked with organizations who’ve taken differing stances on where privacy is a little bit more regulated or it’s a little bit looser. 

I had one where we were looking to promote within EHR, the electronic health record to get to physicians within their clinical workflow. I received a lot of questions from regulatory and from privacy in terms of how does this work? The physician’s being very finite and intimately targeted. 

They’re with a patient. Basically, how are we doing this and making sure that it’s compliant? Really, that’s when you bring in kind of the EHR provider because they’ve set up those privacy regulations. Within the actual EHR platform, whether it’s Epic, Cerner a bunch of others, and they have all those regulations in place, otherwise they wouldn’t be able to be in the business that they’re in. 

99.9% of the companies that we work with in the digital media and promotion space already have privacy regulations in place as part of their platform built in. To be able to pass off to the client side to be able to advertise or promote in those areas.  

Aaron Burnett: Has the shift in privacy regulations resulted in a focus on downstream conversion, that higher value conversion? 

So I asked this question because we work with some clients who are definitely HIPAA-covered entities, and so there are things that we can’t and, maybe don’t track in analytics that we did prior to November of 2022, when that limitation came in into place, what that resulted in is our focus on, okay, if we can’t track the easy thing, the big red, easy button, the lead capture, if we can’t track that in conventional analytics, what’s the moment of conversion that. 

Truly is the value delivery and can we track and optimize that? And for us, that resulted in integration with CRM and tracking events in CRM, right down to new customer start or lifetime value. Have you similarly shifted to track on those higher value but maybe harder to identify initially? 

Todd Weinstein: Conversion moments you always want to track. You always want to know lifetime value of a patient, to be able to understand what your ROI is, as well as your ROA, your return on advertising so that you can say, okay, I’m spending X amount of dollars. And this tactic or this strategy converted a patient, what is their lifetime value in comparison to what I’m spending on that campaign that yielded the conversion? 

And yes, I hear you. In terms of like events within CRM, you can track also on Google analytics. You can have goals or key action fulfillment, and then you have specific digital media that is driving to those key actions. And then what is the value or the worth of the key action fulfillment. 

In comparison to the digital media spend that I’ve invested to get people to that action on the website. So yeah I don’t, I can’t think of anything that, that, and I’m sure there is. I just can’t think of it at the moment. That drastically changed from 2022 when all of this went into place to now it was more about. 

At least in my experiences, a doubling down and greater emphasis on, okay, are we 100% sure that our patients are de-identified with tokenized IDs when they come through to any kind of manufacturer or client facing view? How is that PII handled by specific vendors? How are they trained and how are they identifying those people? 

And de-identifying from a tokenized perspective to be sure that no PII is being divulged at any set. Yeah, so like I’ve gone through kind of privacy audits and compliance, interested in learning more to be able to make sure. That again, nothing is out in the public domain that would allow us to see patient information that we shouldn’t. 

What are the protocols in place to make sure that there’s no data breaches? I know there’s been like large healthcare data breaches in the past with, I feel like it was either on the pharmacy side or maybe the insurance side. Then obviously they do their best to control as well. So typically. It’s more on like the partner, the agency or the publisher side that we’re seeing those protocols in place and then passed off to us. 

Aaron Burnett: Are you putting BAAs in place with partners then?  

Todd Weinstein: Yes. So those are in place to be sure that the left hand is talking to the right and everything is covered from t’s crossed an i’s dotted perspective.  

Aaron Burnett: We have a belief that privacy regulations, while most restricted for healthcare and MedTech today are really coming for every industry. It’s just a matter of time. Do you have a similar or different perspective?  

Todd Weinstein: I think it’s quite possible the closest. Industry that I would compare healthcare to in terms of the level of privacy and compliance is finance, right? So you think about your banking, it’s online, and obviously that sensitive information, and it is held close to the vest in terms of no data breaches. 

I know that there’s a fair amount of this stringent policy and compliance in the finance industry. I don’t know that it will go into other areas to the extent that it is in healthcare. If I think about CPG or maybe e-commerce to me, from what I know, those companies can put out advertising campaigns and other things in the public domain through a much less arduous process than you can in healthcare. I don’t know that will change, but I think the closest comparison I would have is finance for sure.  

Aaron Burnett: What we’ve seen among our clients is that some of them, particularly those who operate internationally, have just decided, you know what? We’re going to work toward a GDPR standard everywhere. 

We’re going to do it in the US even though it doesn’t necessarily apply. And then others who, of course are national and international have decided we are going to comply with the most restrictive posture on privacy, which tends to be at a state level, California regulations. So we will comply with California regulations and we will let that be our guide everywhere we operate, regardless of whether they’re in healthcare or MedTech. We have some e-commerce clients who’ve taken that position as well.  

Todd Weinstein: Yeah it’s valid point, the California Consumer Privacy Act, allowing consumers to opt out of their data being available for purchase. And then obviously if you, if someone’s purchasing it, they’re going to, they’re going to enact certain marketing activities on it. 

I assume I was on the agency side at the time, we were keenly aware of, and we were informing clients in terms of what they need to do to comply with that on a website or more in general. Then, yeah, the GDPR that also impacted website construction and the data collected and then putting together legal language that allowed certain consent or opt-outs to, to occur in terms of the most stringent or the least stringent, that’s more around the company’s posture or stance on mitigating risk and how deep or not they go. 

I think it really comes down to, not the personalities of the folks in the compliance department, but what’s their risk tolerance and how much or little do, are they willing to go with the marketing team? And it’s also dependent on what is the FDA potentially said to companies that are trying to do the same thing or in the same category. 

And if they’re paying attention to competitors that have done something, then it’s not the best idea for us to try and push the envelope if they’re already in that space. And I understand that. But just more in general, right? The compliance and the legal and the regulatory folks are typically very on top of FDA regulations and understanding kind of what flies and what doesn’t. So that when we, in marketing, bring them a piece for approval, it can be a kind of a good conversation to be able to understand where we’ve got to draw the line and where we can, maybe push a bit.  

Audience Targeting in a Restrictive Environment

Aaron Burnett: So we’ve talked about MarTech and measurement and attribution and conversion analysis. Let’s talk about audience targeting. How is audience targeting changed in a more restrictive environment? How are you going about identifying your audiences and then optimizing for performance?  

Todd Weinstein: Yeah, it’s an important one because obviously you want to try and minimize your media wastage as best you can. Again, I would say on the HCP side, easy is not the right word because anytime you’re targeting anybody with promotion, through audience segmentation there’s a lot of nuance involved. 

But with HCPs, again, it’s an NPI number. We know how much or little they’ve written, and then, our goal is to meet them where they are. In terms of the platforms that they visit so that we can serve up information to be able to get them to be interested in, in, the drug, the device, whatever, it’s on the patient side. It certainly changed, Facebook and Instagram have changed a bit in terms of the data that we were able to get access to on the patient side was deeper years ago. That allowed more finite targeting and that is just different now based on regulations that have been put into place. 

Still good platforms, Facebook and Instagram, to be able to promote and typically use, either as a lead gen platform to get them to a destination you want them to be at, whether it’s websites, CRM, signup, whatever it is, or just to use those as in platform advertising opportunities. Yeah I think the other side of is, working with DSPs that have obviously compliance in place and regulatory in place, but also have the ability to use their data to target specific audience members. There’s definitely a data play and if you have the money that can be committed to it, you want to be able to create varied messaging that goes after certain segments of audiences that would cater to what you believe to be their barrier or motivators to getting on the drug or the device. And to your point, that comes back to the data available that tells you about these people to be able to segment them. Then identify them out in the marketplace to serve up tailored messaging.  

Creative Testing and Responsive Ad Platforms

Aaron Burnett: Our experience has been, and in fact the platforms have explicitly said that they’re using creative to target, to identify audiences, and then they’re using creative much more actively to optimize campaigns. 

So our ability as an agency on behalf of our clients to optimize through manual action is significantly diminished. It now is creative variation and ensuring that we’ve got the right data to inform the platforms as you suggest. And then working algorithmically, have you had to change the way that you produce creative or the way that you approach when it used to be AB testing? AB testing is entirely insufficient. Now it’s more like A-B-C-D-E-F-G-H-I-J-K testing all at once and then do that again every couple of weeks. Have you had a similar experience?  

Todd Weinstein: I think it’s become harder yet, and easier at the same time. Harder because of what you just described, right? There’s so many different variations that the creative can take and really if you’re testing more than one thing at a time, you don’t know what’s working and what’s not. So you want to test, ideally one thing at a time to determine whether that did the right thing or not. And then you move on to the next variation of testing. Whether it’s different copy, different imagery, different CTA placement, whatever it is. So yes, that part is difficult. 

But I would say it’s gotten easier because these platforms, some of them have testing and responsive natures built into them. So Google, for instance, with paid search, they now have RSAs, right? Responsive search ads. In the past, you ran with one search ad and it either lived or died, and you knew that based on the bidding process as well as the conversions that you were able to get from GA4.

Now, yeah, you’ve got three plus responsive search ads and it will rotate those and that it serves up the best one based on. What you’ve put into the platform and based on the performance and Meta and Instagram do the same thing. I think where it puts strain on companies is now, you’ve got to invest in content creation and imagery creation at a much deeper level than you have in the past where you lived and died by one campaign. 

So now there’s additional money that goes in. But the argument, and I think rightfully is that, okay, fine, I’m spending more money up front to create the content, but in the long run it’s going to improve my cost per conversion. My conversions in general so that I can figure out what’s working and what’s not, and you can use that across the board whenever you figure out what your best combo is in terms of creative, imagery, copy, et cetera.  

AI Tools and Innovations in Healthcare Marketing

Aaron Burnett: Yeah. So long as you have the infrastructure to capture that data and acknowledge the patterns and action them on a successive basis. Yeah, it makes things only better. What are you working on? In terms of your marketing that you’re really excited about, what is novel? What’s lit you up?  

Todd Weinstein: The evolution of AI is hard to ignore, nor should you be ignoring if you’re a marketer to really help with efficiencies. Operationally, it can just speed to market certainly helps. 

Financially it cuts things for that would take hours into minutes. Just the usage of AI to be able to produce things like copy, even better ways in which to measure and synthesize data that typically took a long period of time is now lightning quick. It’s that integration of AI into everything that we do to make it quicker and more productive. 

I think the other use case would be omnichannel has now been around for some time within the healthcare space, both on the patient and the HT P side. I think it’s probably a little bit easier to implement on the HCP side because it can. Produce alerts and notifications in a very timely manner for the field member to go ahead and make a call on a physician based on certain business rules you set up. 

But there are AI integrations in this too that would basically speed up this process that would be used to say, “Hey, this physician hasn’t been called on in this period of time, or their account size or their market share has shrunk from X to Y. Or they’re now writing competitively in this, with this brand you should call on them.” 

So certain AI driven identification of market movement, good or bad, that then prompts the rep to make a call on that physician based on the data that’s available and provide it to the rep that says, “Hey, here’s all the reasons you should call on this doctor.” I’m seeing AI also being injected into omnichannel to really spark field movement, as well as AI being injected into the segmentation that’s used across patient and HCP for omnichannel that has the ability to change patients and physicians from one segment to another, versus waiting for some sort of quarterly like segmentation hygiene exercise. 

The sooner you can get someone into a different segment based on how they’ve engaged or not with the varied content you’re putting out, the more likely you are to convert them versus keeping them in a static segment and waiting for some sort of manual look at segmentation. That might have someone languishing for 2, 3, 4 months. 

Aaron Burnett: Absolutely. So what AI tools, utilities models, are you seeing and using right now that, that are consistently delivering value?  

Todd Weinstein: Yeah, I think it’s some of the usual suspects, right? ChatGPT is one Co-Pilot through the Microsoft Suite is another, to me. It’s more about how can we put in a business problem, right? 

You’ve got to start with a business problem or challenge into an AI engine and then have it do the work if the business problem or challenge isn’t well articulated or thought out, doesn’t matter what AI tool you’re using because you’re starting with something that’s broken. So I think either of the two that were just mentioned for quick solves and quick queries would be helpful. 

And then I think too, looking at, and I don’t remember the name of the actual AI tool, I don’t, I’m just going to call it Google AI for lack of better terms, but it’s the zero-click searches. Where you have the AI response come up without even having to click on a link. So that is a key one as well, where you’ve got to be able to construct your website in a way that it picks that up with the assumption that even paid media and or organic search isn’t necessarily going to do the trick because you’ve got these zero-click environments where you need to be present. 

The Future of AI: What Stays Human

Aaron Burnett: So when you think about AI and its implications for digital marketing in particular, project forward, gosh, it’s hard to guess a timeframe anymore. You used to be able to ask questions about what do you think will happen in five years with AI? What do you think will happen in 43 days? The world could be entirely different. So let’s project forward, two or three years. What remains the domain of humans? Where are we adding value and how is AI augmenting or replacing work that digital marketers have done in the past?  

Todd Weinstein: I do think, and it’s undeniable, that AI will unfortunately put folks out of the workplace or they’ll need to evolve in a different way to figure out where they stand. I think the human element will always be required, like I said earlier, to be able to crystallize what the business problem is in a way that gets you whatever the best outcome would be when you put that into AI or whatever the AI application is. So for instance, using AI to real time set, to do real time segmentation, whether it’s patient or physician, you’ve still got to put business parameters around what constitutes each of those segments in broad strokes, right? You could hypothetically have HCPs in a couple different camps. The physicians that write us, the physicians that don’t, the physicians that are writing competitors. Then those three categories, you’ve got to define with some business rules, so that if you’re going to use AI to help segment it, knows who to put where and who to adjust in real time. 

I think AI is and will absolutely continue to change the marketing landscape, but there’s the human touch on the inputs for AI to be able to help you solve a business problem and starting off with a business problem. Then also the human touch to probably monitor AI to be sure that it’s going down a path that isn’t too far to the left or the right after it’s been set out to do what you want it to do. 

Aaron Burnett: Yeah, I think that makes sense. I think our thesis is that the highest value work that will be preserved will continue to be the domain of people is deep expertise and the ability to think cross platform. To make connections and to orchestrate rather than be someone with deep expertise in a particular task. 

I think you’re right. You defined a very specific task. Alright. Automated segmentation based on a set of rules that can be understood and I inputted by AI. I think there’s real value there. I think the more you broaden those task sets or even attempt to orchestrate in a workflow, which is very much in the zeitgeist now, the lower quality, the output. 

I have this consistent experience. AI can be so exciting. It can create something almost good, so fast, and the almost good is an illusion. It’s very exciting at first, and then you realize, oh, actually the last 10% to make this good takes about 80% of the time it would’ve taken to do this if I’d just done it. It’s just a different way to get to that outcome.  

Todd Weinstein: I think you’re right. I think AI can do the bulk of the heavy lifting, but. I’ve received emails before where I’ll read it and I’ll just say this is AI with zero human element put in, and they just took it from AI and here it is in front of me. 

I’ve also spoken with people in different settings where I’ve gotten the impression that they were using AI to ingest what I’m saying and then spit out a response that they are using a computer screen to read off of. When you’re doing that kind of thing in such a quick rapid fire change just like this, let’s say I was using AI to respond to your questions without time to reflect and likely change it and humanize it, it comes off as inauthentic. And so I think that’s the part that, whether it’s healthcare or not, folks just need to be mindful of and careful of so that they’re. They’re putting their handprint on it to make sure that it’s got some human touch.  

Aaron Burnett: Yeah, absolutely. We are in for some interesting years ahead.  

Todd Weinstein: For sure. 

Aaron Burnett: I’ve really enjoyed talking with you. I appreciate you making the time.  

Todd Weinstein: Likewise. Thank you. Thanks for having me. 

Sponsored by Wheelhouse DMG