How to Layer 3P Data and Reach Ideal Patients

Announcer: Welcome to the Ignite Podcast, the only healthcare marketing podcast that digs into the digital strategies and tactics that help you accelerate growth. Each week, Cardinal’s experts explore innovative ways to build your digital presence and attract more patients. Buckle up for another episode of Ignite.

Lauren Leone: Hello, everybody. Welcome back to Ignite Healthcare Marketing Podcast, back with Rich Briddock, our Chief Strategy Officer. Today, we’re going to be talking about third-party audiences, which is something, Rich, you and I talk about probably daily at this point. The point of third-party audiences is to expand reach outside of what’s available in platform. A lot of our clients nowadays are using them to get more targeted, to get more focused, to tackle the lead quantity with lead quality issues that they might be having, and to ease operational concerns with high volume of phone calls. Rich, I want to start with a little bit about the use cases for third-party audiences, then we can get into where to find them, what are the right tech and tools, where can you plug them in? Start with a high-level overview for us.

Rich Briddock: First of all, thank you for having me back on Ignite.

Lauren: I finally let you get to work.

Rich: Yes. It’s great to be back in my favorite chair, sitting next to my fireplace that doesn’t put out any heat. I think in terms of, you sort of described the use case and the need pretty succinctly in your intro, but obviously in healthcare in particular, but just the way that platforms have been going with Meta, there’s been this shift towards restricting audiences that you can target and more of a shift towards an algorithmically driven advantage plus type scenario, especially on Meta and also on TikTok, where essentially you’re just targeting a broad audience and those platforms saying, “Don’t worry about trying to dial the knobs in. We’re going to find the people who best suit you.”

Which works well for some businesses, especially if you’ve got enough volume of conversions and you can run the monies where the algorithm can really learn, and the audience is broad enough and there’s enough of those consumers. If you’re trying to target a very specific audience, or if you don’t have the signals, like the qualified signals to feed that algorithm, to help it find the right people, the inability to really dial in on who your audience is in platform natively is becoming a huge problem. That’s where these third-party audiences come in. The main use case that we see in healthcare is targeting by two things. It’s by the clinical-based targeting, people who have the right diagnosis, or people have had the right procedures, or are taking the relevant medications.

Then the other is the payer layer. Are you commercially insured? Are you commercially insured by a specific payer, like you have Blue Cross, Blue Shield, you have Aetna, et cetera? That’s really important for some of our clients who have a limited amount of payer relationships or different payer relationships by state. Insurance really is make or break for their services. They can’t treat Medicaid, or no one would ever do self-pay. You’ve got to have the right payer mix in order to be a legitimate, qualified leader, essentially.

Lauren: Taking a half step back of why this matters so much for healthcare versus other industries is other industries can use their first-party data. That’s a huge limitation.

Rich: Yes. Yes. You can use first-party data, you can build lookalikes from your first-party data in other industries. There are solutions out there who are trying to enable that in healthcare too, like CDPs, but obviously it depends on what your compliance stand is.

Lauren: Still a huge gray area.

Rich: Massive gray area. Obviously, every organization has a slightly different stance on that. It’s not to your point, the same tried and true tactic that you can deploy. Solutions like Google actually will just not let you leverage your first-party data. Meta is moving that way if you’re partially restricted or fully restricted as a healthcare advertiser on Meta. You also can’t use first-party data to build audiences now. Yes. These third-party audiences help you to really fill that gap. Oftentimes I think it’s not just the restrictions on the first-party data, it’s do you even have that good first-party data that’s easily accessible to you, where even if you didn’t have the restriction, you could quickly get hold of it and you could leverage it for segmentation of targeting purposes.? Yes, that’s it in a nutshell in terms of solving for the problem statement.

Interviewer: If we’re thinking about one more baseline background setting statement, what can you target in healthcare without a third party? What’s the standard if you’re not using third party? I know you mentioned Advantage Plus, which takes into account what?

Rich: Advantage Plus is, it’s essentially Meta’s version of a propensity model. They are looking at who is converting, they’re gathering data in the learning phase, they’re looking at who’s converting, and then they’re basically finding the correlation of various traits and saying– similar to a lookalike audience, and then they’re saying, “Okay, who looks like this? Who do I think is going to behave like this? Who’s looking at the same things that my converters are looking at online? Who fits the same demographic profile?” They’re looking at hundreds of thousands of data points that they have.

Lauren: However, those data points include no health context. That’s the big difference.

Rich: No health context, or at least no health context that you’ll ever know about, right? They might have the health context in terms of, they know that I’m browsing on a certain site, and they know that my converters have also browsed on that site. Therefore, they can put that together and say, “Okay, yes, people who look at this site have a much higher propensity to convert, so we should target those folks.” That might include health context, it’s just that we’ll never know what a health context is. You’re running blind, like you’ll know if Advantage Plus does or does not work, but you will not know why. You will not know which audience segments inside of Advantage Plus are working versus which are not working. You’re always operating blindly and you’re just trusting best.

Lauren: Which is okay. Yes, we’ll talk a little bit about comparing the two in a minute, but jumping into the details of the third-party audiences, talk to me about direct versus model versus lookalike. What are the different options that we have in healthcare?

Rich: Different providers have different approaches. There are deterministic providers, which people in those audiences actually have that trait that you’re targeting. It might be that you have this condition, you might have been diagnosed with anxiety, as an example, or you might have had a specific procedure, you may have had a hip replacement. You definitely have Aetna, Blue Cross, Blue Shield, that is a deterministic audience within a high degree of probability that is accurate, at least on the payer side, it’s a high degree of probability. On the actual condition procedure side, you’re in the claims data that has happened to you.

Lauren: The claims data is what’s tested.

Rich: It’s the claims data.

Lauren: That is HIPAA compliant? I get asked that–

Rich: That is HIPAA compliant. It’s tokenized, then it’s pushed to LiveRamp or Throttle, and then it’s pushed into an ad platform. No one can identify who that user is because it’s tokenized before anything is pushed through for audience space targeting. Beyond that, and obviously the advantages of deterministic is, you know you’re reaching the person who’s actually got that condition or has had that procedure or taken that medication. That’s the smallest, the sweet spot. It’s right around the bullseye. The challenge with claims-based audiences are, it may be too late. By the time they turn up in the claims data, it may be too late to reach them, depending on what you’re trying to reach them on. To go back to your previous point, there are some segments inside of claims data that are considered sensitive, and you can’t ship those audiences.

Lauren: What are they? Just for people listening?

Rich: Addiction is one, pregnancy, fertility is another one. Those are the two main ones that we know.

Lauren: Teen and adolescent. Some behavioral health.

Rich: Yes. You can’t pediatric, you can’t do. Strange ones that you might expect to be sensitive, like oncology, you actually can’t target. Anxiety, depression as well, some of the behavioral health stuff you can target. It’s not necessarily just all the categories that you think that would fall into that most sensitive category. The other challenge with deterministic audiences are, even though it’s your bullseye, those audiences could be very small. Even though these third-party audiences tools have all these different mechanisms to help improve your match rate when you’re shipping them to ad platforms, it could still end up being a very small audience target against your frequency could get very high. There’s another–

Lauren: Recommend keeping your geography fair. If you have the ability as an organization to, in a multi-local environment, activate that layer at a national level or regional level, it’s not going to be layered with geography, is the challenge.

Rich: Yes. The highest possible reach, I would add those audiences. Then what you have is you have another group of audience providers that are probabilistic. They are modeled audiences where they take a seed audience, which is that deterministic audience that we were just talking about from the claims data or from other sources. Then they use propensity models to basically say, “Lauren, you look like someone who may have this diagnosis, who may be in the market for this type of procedure because you have a lot of other common traits with those audiences.”

We’ll put you in a likelihood decile of needing that procedure. It’s a way of saying, based on a decile approach, how similar are you in terms of all these other factors that I’m looking at to people who have had that procedure or who have that condition in terms of your behaviors, your social determinants of health, and all of these other things used to build that model. Then that obviously expands your reach out massively, creates a huge audience. Then you can target down those deciles. What these providers allow you to do is to say, let’s say 10 is the highest decile. You’re the most like those people who are on the seed list. I could then target deciles 8, 9, and 10.

Those are that next concentric ring around my sweet spot, my deterministic audience, that still highly likely that these people would need treatment based on these traits. It gives me the opportunity to scale. It also gives me the opportunity to get ahead of people who have not had that diagnosis yet or who have not had that procedure yet.

Lauren: Maybe researching contextual, going places physically in the real world, yes, shopping certain things.

Rich: I think one of the examples that we’re going to talk about later is PT or a mammogram. We have a client who offers mammograms. Given that they offer mammograms, they don’t want to target people who have just had a mammogram. They might want to use deterministic audiences to target people who have previously had a mammogram, maybe two, three years ago, and have not had one since, and therefore they’re lapsed and they should get another mammogram. They’re not going to want to target a deterministic audience of someone who’s had a mammogram in the last 12 months because they’ve already had a mammogram. They would use a modeled audience of people who likely fit the demographic or fit the pensity model to need a mammogram, and then actually use the deterministic audience to exclude those users, to suppress them of anybody who’s had a mammogram in the last 12 months. That’s a use case where you’re using the model audience and the deterministic audience together to try and hit that sweet spot of people who are likely in the market but have not yet had it.

Lauren: Anything else deterministic, probabilistic, you want to talk through before we talk about where can we activate these?

Rich: There’s other audiences that I would say that we work with where it’s more contextual-based. Obviously, anybody can go out there and do contextual targeting on programmatic, and you can do it through Google as well. We have other providers who are taking search-driven behavior across a much wider ecosystem. They’re basically capturing search data from various sources. Their whole goal is to allow you to reach the other 94% that don’t click on your Google search ads.

Lauren: They’ve searched the term.

Rich: They’ve searched the term, they’re in market, they’re doing the research, and they’re building an audience that then you can target on Meta, you can target on programmatic, you can target digital audio. Whereas typical contextual, you’re going to be targeting a small subset of those through your own contextual efforts, where you’re not going to be able to deploy that audience across social platforms. You’re limited to deploying it across display, native, maybe some OTT, and digital audio. These solutions, because they’re essentially creating an audience and they’re matching it to an identity graph and device IDs, it allows you to reach that contextual audience across pretty much any image.

Lauren: Everywhere else that they could go.

Rich: They could go, yes.

Lauren: I think that’s a good segue. Where can we activate these audiences, and where can we not?

Rich: Anywhere. Because you’re pushing them via LiveRamp or you’re pushing them via Throttle, you can activate anywhere that can ingest data audiences from LiveRamp. The most common are going to be Meta, TikTok, your DSP, programmatic partners. That will be agnostic. If you use TradeDesk, if you use DV360, if you use Viant, Illumin, any of these providers, you’ll be able to push those audiences too. Where you will not be able to push these audiences to is Google. If you’re a healthcare advertiser, you just cannot push third-party audiences into Google to target, unfortunately. I have heard some rumors that that might change at some point, but it’s just not a capability. You have to.

Lauren: Never know until it’s there. Not with Google. We’ve walked through a couple of examples. We know the details, the logistics, the cost side of things. Obviously bringing in a third-party audience is going to come with, I assume, a CPM markup, and they’re going to have varying degrees of cost tied to them. At the end of the day, how do you recommend actually deploying this in a strategy? When should you use it? How do you know if the cost is worth the result?

Rich: Different platforms that you deploy on have different mechanisms of utilizing the cost. Not to get too deep into it, but if you push these audiences to Meta, you will not be able to bake your cost into Meta. You basically have to track how many impressions you’re serving because it’s all done on a CPM basis. You will know if I serve 100,000 impressions and I’m paying $5 a CPM, then I have to pay $500 to my data provider in addition to what I spent on Meta. You’ve always got to be aware that there are incremental costs that you’re incurring by using these audiences. If you run through a DSP partner, oftentimes you can do third-party billing, which is where those costs are baked in, actually into the platform. The cost that you’re deriving from the platform, the third-party audience costs are baked in. You’re getting a more holistic view.

Lauren: Essentially, entering it on the backend and saying, this is how much I’m paying to use this.

Rich: Yes. I’m shipping this audience. This is how much, and then it takes that into account as you run against it, which is obviously much more helpful for managing it. [crosstalk] I’d say some of these platforms have no minimums. It’s just, you come in, you run an audience CPM fee. Some other platforms, especially where there’s a BI component involved, which we didn’t talk about, but a lot of these audience providers have a BI component where you can actually go into a Tableau interface or go into a reporting interface and identify where those users are and how big those audiences are before you ship them. They will have like a minimum spend commit that you need to stick to.

If you’re not running it through an agency that already has a relationship with audience providers, then just be aware that you might have a contractual agreement that says I have to spend $20,000 a year or $30,000 a year against these audiences as part of my license for this platform and access to these audiences.

Lauren: One more use case for these, which I think we’re going to do a whole other episode on. Everybody listening, tune in for that. It’s coming up, the provider side of the equation. Just one sentence and we’ll leave it to a whole nother episode.

Rich: Most of these audience groups also target providers. The HCP targeting is obviously significantly more expensive. It’s usually between $15 and $25 CPM, but they have some cool stuff in terms of the ability to target providers based on their referral behavior, based on the types of prescriptions that they write, based on the types of conditions that they treat. The goal with that is you’re targeting the right providers, the really high-value providers. Even though it’s a higher CPM, obviously, the lifetime value of flipping a high-value provider that will give you 100 referrals a year, 200 referrals a year, the lifetime value of that is huge. Reaching them and engaging with them, and helping them to refer to your business is massive.

Lauren: We’ll spend more time on NPI and NPI to DTC and some other concepts here. Closing, people interested in this, vendors that we use. Can you shout out the names?

Rich: Yes. PurpleLab on the deterministic side. Swoop, Definitive Healthcare on the model side. Then V-Analytics is the contextual partner that we’re working with. If you’ve got an audience that’s hard to reach, like pediatrics or ABA, V is a great one to leverage for that.

Lauren: Awesome. Thanks, Rich. We’ll be back for more on third-party audiences and the HCP side. Thank you all for listening. Subscribe, follow us wherever you listen, and we’ll see you guys next time.

Announcer: Thanks for listening to this episode of Ignite. Interested in keeping up with the latest trends in healthcare marketing? Subscribe to our podcast and leave a rating and review. For more healthcare marketing tips, visit our blog at cardinaldigitalmarketing.com.