How to Drive Real Results When Your Data Isn’t Perfect

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: Welcome back to Ignite: Healthcare Marketing podcast. Today I have Scott Cameron, one of our group media directors here. Scott, we’re going to talk about– We love to host episodes about perfect measurement and what our ideal dream state is, but what we don’t often talk about is what to do when that is just not your reality right now, or you’re working towards it, but you’re on a roadmap that’s going to take 12 months. Today, I’m excited to get some tips and tricks from you on the flying blind or flying partially blind conversation that we have more often than not.

Just to set the stage here, ideal destination in terms of tracking is something we kick off every partnership with. We want the clients to understand what’s possible, what an ideal state would be, and then we essentially work backwards to where they are today and figure out how to take baby steps towards that. Oftentimes, we have to deal with imperfect data. Setting the stage, what are some of the challenges you often face? The honest situation about where most healthcare groups are today and what they’re measuring.

Scott Cameron: Honestly, the real answer is that it’s very rare that somebody has everything completely buttoned up. There’s usually something missing, whether it’s we’re not tracking calls all the way through to a booked appointment or we have some level of drop-off or attribution challenges. There’s usually a mix of that happening. I actually love the topic for this one because it really focuses on what are the practical things we can actually do to improve the algorithm, improve performance, while we’re working to get to that latter stage of, “Okay, everything looks good, we’re tracking and confident that we have everything exactly buttoned up.”

Lauren: You mentioned it. You have everything from, I’m tracking clicks to call to I have some data on calls, but I don’t know what happens afterwards, or I know a form was submitted, but how do I know if it was a new patient versus returning? Then how do I build a digital campaign that optimizes as close as I can or leans into the things that I care about the most, knowing I just don’t have all the perfect answers? We’re going to talk about the specific examples. If you think about a ladder framework here and building upon it and climbing it and getting to the ideal state, there’s an evolution and the steps that we can get through.

When you think about wrong number one, and oftentimes a lot of healthcare, we know people still book through the phone. Let’s talk phones first, and then we can get into some of the other mechanisms. With calls, what is the spectrum that you see? What can people work towards if they don’t have anything today, and then what’s an ideal state?

Scott: I’ll start with ideal state, and we’ll work backwards. Ideal state is we are able to track those booked appointments, and we’re able to segregate by new patient, existing patient. That’s dream state, is, we’re able to track all that through. Often what we have-

Lauren: I feel [unintelligible 00:02:53] have good stuff, right?

Scott: Yes, and anything beyond that is even better. That’s ideal state. I think generally what we find is, even if somebody is using a call tracking partner, they’re not necessarily at that stage. One of the really good ways that we can approach this and still give meaningful information to the algorithm, which is crucial, the more we are giving the algorithm important signals that actually lead to business outcomes, the better off we’re going to be, the better off performance is going to be tied to those outcomes.

At a basic level, one of the things we always look at is something like call scoring. Looking at a specific duration of a call instead of just all calls. A call that’s, let’s just say, 90 seconds is a random example. It is going to likely be better than a 45-second call. The really interesting thing about this, and it ties into other things around ops, but the really interesting thing about this is it’s a pretty quick conversation with your ops team of, “Hey, what do we really see in terms of a duration when there is a quality call?” Because that’s always going to differ, and even better if you can have it, is connected call duration.

Someone actually picks up the phone. How long does it typically take for someone to actually book an appointment? That is a quality lead. Leaning on your teams internally, having those quick conversations, and then really just setting that up within Google Ads. It’s not a difficult thing to do, but it does improve our overall quality if we can’t get that data like a booked appointment.

Lauren: Moving from something like all calls count and all calls, are calls that may or may not have reached a human being, left a voicemail, came after hours. All these things to a call duration is a baby step in the right direction. I know there’s– Again, every call tracking solution is different. There’s post call, key-press options, and ways that you can integrate with your call center, and some other things you can do. Tilt directionally, understand what’s happening.

Then, I think, the ultimate products nowadays have AI listening, and that’s where we’re hoping clients can get to, is at least using AI to determine what is the likely outcome of this call. Even if it’s 80, 90% accuracy, it’s better than where you were before, which is just all calls matter.

Scott: Yes, that’s exactly right.

Lauren: Then the ideal end state, like you said, is if we have a true feedback loop, AKA, that call is pushing into a CRM or an EHR or some data lake where you’re actually matching and verifying that’s a patient that was on a provider’s schedule. That would be where we would love to go with calls.

Scott: Yes, exactly.

Lauren: You think about calls, and then you’ve got forms. What’s the situation there? Where are most groups at? Because online scheduling isn’t always available or groups maybe haven’t adopted it yet, or they’ve got 500 brands, and so rolling it out is just a year-long process. What are you used to seeing on that front?

Scott: It’s exactly that. It’s such a mixed bag. Sometimes you have large dental practices that maybe have 200 different locations, and all of them have done it differently over time. That’s probably one of our bigger challenges of how do we make sure that this is consistent? Then how do we work with all of these different types of online schedulers? Maybe some of them have a manual form submit option on their website. Working within those, that’s definitely a challenging thing, especially when you have some of those multi-locations like that.

If we can’t get to that end state of, we have a booked appointment, and even further beyond, did they show for their appointment, and do we have a real revenue opportunity? Here is some micro conversions we talk about a lot. How do we look at this when you can’t get that full visibility into a form submit and booked appointment? Are there certain things that are happening prior to that that we can truly track and assign value to? Maybe it’s like an insurance verification page, for example, or they’re selecting the date of the appointment. All of these things theoretically show intent and purpose into booking an appointment.

If we can assign values to those, that’s one thing we love to do as well. Again, if we don’t have that ideal state, it’s, hey, how do we say this is worth a certain amount, and then teach the algorithm, “These are the types of events we truly care about that have meaning,” and ultimately lead to the outcome that we want.

Lauren: You think about a standard form, and is there maybe a field or two in that form that if they select a certain thing from a dropdown, or enter a certain bit of information, or actually fill out that field, it’s maybe like a plus-up indicator that that’s a little bit more valuable of a lead than the average person filling it out. Can we collect that information upon submission? Not the PII information, but the like, “Yes, it was submitted with this information,” and we can give that a higher value.

Scott: That’s exactly it. There’s a checkbox that’s like are you a new patient or are you an existing patient? That’s the type of information we would try to pass. Again, avoiding any PHI, that’s obviously what we aren’t trying to do, but it also avoids any of those compliance conversations, and that hold up in that way too.

Lauren: Then we’ve got the online scheduling. I don’t want to call it a conundrum because we love it, but the easiest way to deploy online scheduling is typically to work with a platform where they’re clicking off and scheduling in some third-party URL. Then you have this challenge of, I got to click into a scheduler, but did I get anything else from there? What are some of the workarounds you’ve seen in that front regard?

Scott: That’s really tricky, too, that we often have seen where it is a click into the actual scheduler. That’s obviously not an ideal state because there could be some level of drop off there. I think any of those micro conversions that are leading up to those events, multi-page sessions, for example, that would show intent, those are the types of things that we’d want to be tracking and assigning value to in those situations, in addition to, okay, then they click into the event. Stacking those things and assigning some value as we go.

Lauren: One of the tips I love to give people, and it’s not always possible, but is to at least ask the organization that the tech you’re using for online scheduling, whether after it’s submitted, maybe there’s a five-second delay, and then it redirects back to a URL, and perhaps you can capture just a hit on a thank you URL. That’s always something we like to toss out there, too, as a nice to have, if it’s possible.

Then you’ve got this whole other world of manual feedback loops. Talk to me about how that is additive to some of the things that we just talked about in terms of what we can and can’t pixel, what we can and can’t pass back.

Scott: I want to say one of the simpler things that you can do that is really effective is just making sure you’re having that conversations with the operations team and understanding the quality of the leads that come in. It’s so basic in the sense of you’re just having this conversation and saying, these leads that are coming in, are they quality? The step beyond that, I would say, is that you are taking that information and then importing it back into the platforms with some system. Maybe they score it on the back end and say, ” These leads that did come through were good quality leads,” or “They did lead to a book deployment.” We can pass those back into the platform and then leverage the algorithm for those.

Lauren: Got someone on a front desk, so taking 30 minutes every Friday to highlight in a sheet which ones were good?

Scott: Yes, it’s exactly that. That’s why I say it’s so simple because it sounds like-

Lauren: It’s so much human power.

Scott: Yes, and it’s not something that we necessarily want to do to have this manual process. It’s better if it’s automated. It’s better if it automatically scores all this stuff and passes back in naturally. That’s all ideal state. Obviously, we’re not talking about ideal state here. We’re talking about that stage before where we’re working on it. This is a good gap where we could say, all right, how do we really get that feedback on quality of leads and quality of appointments and all of those things just manually. Like you said, it could be someone at the front desk who’s just saying, “Hey, this call came in, it was good quality, and we can tie that back.” Or “This form submit came in, what was it? Was it worth a book deployment or was it spam or something?” Just making sure that we’re getting that feedback loop with ops is incredibly important.

Lauren: Yes. We’ve talked about some of the workarounds, the things that groups can do, don’t just give up and say, “I’m going to track clicks and that’s the end state of where I am until I’m at perfect.” Groups working on taking these steps, then there’s all the platform hygiene that comes with every time you go and say, “I’m going to change the signal. Every time I’m going to push new information, what happens to the old information? How do I think about my campaigns evolving? How do you think about that process and what really goes into it?

Scott: That’s obviously incredibly important that we have that foundation. I’ve heard this analogy once, you have to fix the pipes before you increase the water pressure. You don’t want to just flood this stuff and then not be able to actually to pull it off. A couple of the things that we look at are things like, within the platforms, enhanced conversions, for example, was a really good one. CAPI, when it comes to Meta, just making sure that we have the best data matching that we can within the platforms and that we’re doing the things that we can to improve that. Both of those are impactful in the sense of we are going to get more data as well, which then impacts the way that the algorithm is able to perform.

Then obviously, one of the other big things that we focus on are just looking at the conversions that we are optimizing to. One of the things we see very commonly is we might have 10 primary events and we’re really just telling the algorithm all of this information that it doesn’t really fully understand what we care about. What we want to do is teach the algorithm all the time with all of these different things to get it to a point where it’s operating and doing the things that we want it to.

If we give 10 signals and say all of these are important to us, it doesn’t really know how to act. One of the things you do from a platform hygiene perspective is audit the account and say, are we duplicating things? Are we teaching the algorithm that a click is the same value to us as somebody who has visited the site multiple sessions and gone to the insurance verification page? It’s probably not. We want to make sure that the algorithm understands the value of those events and it’s cleaner. We want to make sure we’re not doing duplicate efforts.

Lauren: Yes, I see the mistake. Oftentimes we talked about calls and let’s say currently you’re tracking all calls and you have a great operations conversation and you say, “Look, 180-second calls are going to be five times more likely to be a new patient than all calls. It’s not perfect. We know that, but it’s directionally moving us towards where we want to go.” You introduce and start sending the 180-second call signal, but you keep all calls also as a primary. Then all of a sudden, every call that was both a call and a 180-second call is getting double counted, but you’re all calls are still being counted, and you’ve got this dynamic where you have way too much going on and still no clarity on what matters.

You talked about primary conversion signals and you can set things to secondary and you can observe what happened to an old conversion signal, but you don’t have to be optimizing towards it. Definitely, I think that platform hygiene is one of the things that, unfortunately, gets missed, even though people are working through this evolution. I’m glad that hygiene is on this four-runged approach of what we need people to focus on.

Scott, if could leave our listeners of this episode with one practical piece of advice on where to start, because I think this is a very overwhelming topic to people, especially when you’ve got compliance involved in legal and you just feel frozen or stuck. What would be one of your starting points?

Scott: Yes. I think one of the big starting points is just reviewing your account and looking at what you’re currently doing and then taking some of the advice that we said earlier on these different rungs of, okay, if it’s a form, then here’s the things that you can do. If it’s a call, maybe it’s call scoring. Just auditing what your existing state is and saying, do we have that hygiene in play? Where are our gaps? Where can we make little incremental changes to what we’re doing in monitoring and tracking? How will that have an impact overall? It really is one of those things where you can make progress without it being perfect and you can really lean into some of the advice that we gave on this lovely little podcast.

Lauren: Yes. Progress over perfection. I hear that a lot. I think that’s the mindset we all have to have right now because we work in a very complicated landscape. Thanks for joining us, Scott. For all of you listening, like, share, subscribe wherever you get your podcasts. We hope to see you guys next time.

Scott: Thanks for having me.

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.