How We Fixed Meta’s Lead Problem

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, Cardinals experts explore innovative ways to build your digital presence and attract more patients. Buckle up for another episode of Ignite.

Rithika Rajesh: Hi, everyone. I’m Rithika Rajesh, and I’m the head of biz dev here at Cardinal. I’m here to introduce Lisa Fisher and Caro Costa, who lead our paid media, social, and programmatic team at Cardinal. They’re going to be speaking about how we fixed Meta’s lead problem.

Lisa Fisher: Hey, everybody. We’re so excited to be here. Caro and I have been putting in a lot of hard work to bring this forward for you all. As mentioned, we do lead up the social and programmatic team here at Cardinal. Even the fact that this pod that we lead exists is a testament to the fact that paid, social, and programmatic are just becoming so important for marketers in the healthcare space, and increasingly so within the past couple of years to warrant the fact that we need to have our own pod now.

We wanted to dive into the perception problem that we really are seeing when it comes to Meta, where advertisers think that Meta is not meant for lower funnel objectives and it’s only meant for upper funnel, but we’ve actually found that it can be a really effective lower funnel tool. We’re going to dive into that.

We’re also going to talk about what’s changed. Meta as a platform has existed for decades now, which is crazy to say, but there have been a lot of developments over the years, a lot of changes from a technical capability standpoint that now allow us to deliver really strong lead campaigns on Meta. We’re going to go into what’s changed, the latest technology that you need to be aware of.

The solution is how Cardinal cracked it. We’ve done a lot of test and learn. We’ve had a lot of various advertisers that were running Meta on that we’ve tested different scenarios. We’ve really come down to a pretty refined science on what that social playbook needs to look like for a healthcare marketer, and ultimately how you’re going to be engaging users in the most efficient full funnel strategy on Meta and driving leads.

We have one specific case study out of many that we’re going to highlight here to show you some exact numbers and then overall discussing the shift where Meta is going to be in 2025, 2026, and beyond. Meta gets a bad rap in healthcare. I think also in general, advertising as a whole, like I was saying, it’s really viewed as a strong upper funnel UGC platform, which it still is and it always will be. That’s where it got its roots.

If you watch reels or go on Instagram or TikTok, all of social media has its base in that grassroots user engagement. How do we pivot that to also be a vessel for driving qualified leads? Overall, Meta lead forms are a tool that we can utilize, but it has to be utilized the right way and intelligently when it comes to targeting and the overall setup.

We want to drive lead quality that is strong, not just a huge volume of leads that are not qualified. We want to make sure that there is qualification within the lead form itself to make users who are engaging with the form express to various degrees that they are the right audience to be receiving this ad, and then be following up with the actual lead form to those users specifically, so that we’re eliminating any waste.

We’re making sure we’re really focusing on the most qualified audiences. We need to basically be pre-filtering our leads, including various questions to make sure that they are the right audience for the specific product or service that we’re marketing, including things like insurance verification or specific insurance carriers. Overall, really, there are several tools that we can use to get high-quality, efficient leads within Meta.

It’s just all a matter of how it’s being used and how the lead campaigns are being implemented. As a whole, Meta has grown up significantly over the years. We always think of Meta as running awareness campaigns, running UGC, brand awareness, brand loyalty, which it can and should still be used for, but Meta has silently, in the background, been rolling out a lot of new technology to compete with some of the bigger players in the space like Google.

I think they’ve really got an eye on other technologies and platforms that are employing AI, employing a lot of in-platform seamless optimization tools that a person doesn’t need to be managing, but that utilizes AI learnings. Meta is constantly developing these new tools. Meanwhile, a lot of marketers are still setting up campaigns the same way that they used to and still thinking about Meta in the same way.

Some of these new tools include SMS verification, which is where you can ask a user who is filling out a form to enter their phone number, or they can also do a CAPTCHA, where they have to prove that they’re a human. This eliminates the spam that would come through in the lead. It may decrease overall volume, but it’s going to make sure that the users who are getting through are qualified and ready to convert.

We also have lead forms that can utilize conditional questions, which Caro is going to get to in a minute, making sure that our users are qualified. Then CRM integrations. This one is huge. The backend of Meta, they’ve been pushing CAPI, which is conversions API, which allows for greater transparency between the client CRM data of actual patients feeding back into Meta.

This is going to allow us to target in a smarter way based on the actual patient attributes that are coming through the actual patients who are coming through in the CRM. What is the solution? We have a new strategy that we’ve really implemented and a lot of tips and tricks that we’ve been optimizing across various clients throughout the past couple of years. We’ve really learned to leverage client CRM data, as I mentioned in the previous slide, to make sure that we’re capturing users most likely to convert.

We’ve also learned how to use we have essentially rebuilt our entire funnel strategy on paid social to make sure that we are ingesting all of the correct CRM data and using it intelligently. We use the data pass back from the CRM integration to make sure that we’re hitting the right users at the right time with awareness messaging, then retargeting awareness video viewers in the middle of the funnel to drive traffic to the site, traffic to engage with different creative units that we’re running in Meta.

Then ultimately making sure that the audiences from each stage of the funnel are being nurtured and passed down to the next stage of the funnel, so that we have only the most qualified ready to convert leads at the bottom of the funnel. That’s going to eliminate cost. It’s going to eliminate redundancy. It’s going to make sure that you’re targeting the most qualified audiences effectively.

Moving into our case study, for this specific example, we had a client that we cut the cost per new member by 73%. That is the legitimate number. It was a very strong turnaround. It happened in only a matter of months. It was from implementing several of these optimizations that I’ve mentioned that Caro was going to get into. Now we’re at the point where these campaigns are running pretty effortlessly.

They’re lean, mean, optimizing machines with minimal in-person platform adjustments. We’ve gotten to the point where they’re smart and they’re able to go out into the world and continue to drive efficient leads. Working into the case study section, there are several different considerations for how to structure your Meta campaign appropriately to drive the most efficient leads.

Caro was going to get into this next. The first stage is the creative messaging. Testing a variety of creative types and different promotions and offers is going to be one of the first things that you need to consider. Obviously, Meta is similar to Google in that it likes to have different variations so that each individual user is not all the same and that there are different variations of creative that can be served to different people depending on who they are.

We also then push qualified audience signals. This means retargeting users who may have seen an ad on another platform. We’ve done some retargeting with programmatic users who’ve seen CTV. Also, retargeting people who viewed a video in an earlier stage of a Meta campaign, as I mentioned earlier. Third is filtering spam. That’s just the lead quality control. Fourth is analyzing the lead quality coming in from the form, identifying who is coming in from the form. Are they converting? If not, how can we eliminate those people? If they are, then how can we get more of them? Then ultimately driving to the lower cost-per-new patient target goal and optimizing from there. With that, I’m going to pass it off to Caro to walk us through the case study.

Caro Costa: Awesome. We’re going to go through each of the pillars that Lisa just described and really give you the tools and the examples of how we tackled each one of them just to put a real case study in front of you to make the learnings a little bit more concrete. The first pillar was identifying messaging strategy. We always know that we want to deliver the right messaging at the right time to the right users.

That’s a common phrase that we use as marketers. In this specific case, we’re talking about we need to identify what really gets the user to submit the form. We want to identify what drives them to submit it and click, and leave their information. In this case, for this specific client, you see that we had two different tests. The first test was really focused on testing whether users were more engaged or were more likely to submit the form.

If we talked to them about the results of getting treatment or giving them words of encouragement, encouraged them to actually begin the treatment. Whether focusing on how easy the process was or how their life was going to look like after. What we found is actually ads that were focused on encouragement had a 25% lower CPL.

The second test is focused around an insight that the marketing team gave us that basically meant that people who had insurance were actually more likely to become patients than people who did not. Obviously, I’m guessing this is the case for most of the healthcare advertisers. We actually ran an A/B test too, mentioning the insurance coverage on the ad.

We also found out that insurance-based ads, had an 11% lower CPL. Why this is important? It’s because when we get to the stage of optimizing the form and to what Lisa was saying, going really for the quality, we need to make sure that our dollars are going towards those creatives that really move the needle from a conversion rate standpoint. We need to filter out and learn first what resonates and moves the needle from a lead from submission volume standpoint.

As we move to the second pillar, this is also related to audience. We also want to make sure that at the action stage, we’re fueling Meta the right signals. We don’t want to just serve our impressions to everyone. We want to serve impressions to audience who have at least interacted with us at some point to actually point Meta in the right direction. Why this is important? It’s because it would also set us up for success when we need to scale.

First, we want to prove social value, but we want to show our ads in front of the right people. Once we do that would also be the key. Having these remarketing pools will be the key to scaling later. If we have a video views remarketing audience, we can later on expand the remarketing window. If we are remarketing cross-channel, meaning from programmatic, we can also then expand the remarketing window, expand to third-party vendors within programmatic, and have a wider pool.

In this case, for example, when we added video views remarketing into our action stage in bottom funnel, we were able to triplify spend while also improving our lead quality in this case. 24% of total leads were marked as qualified versus 17 before without adding this video viewers remarketing audience. Now, at this point within this specific case study, 24% of total leads being marked as qualified was still a low percentage volume of qualified leads.

That’s when we move to our third pillar, where lead quality becomes our main focus for optimizations. Here’s where conditional questions come into play. In order to identify what was the main reason of the fact that only 24% of our leads were marked as qualified, we had to look into what was the number one reason why our leads were marked as unqualified?

As you can see in the bottom chart, in March, 57% of our leads were marked as not a lead/spam. What did we do to improve this percentage by April? As you can see, that number decreased to only 19%. This is where conditional questions came into play. Basically, conditional questions would be asking the users, in which type of treatment are they interested?

Are they interested in in-person, virtual treatment? Are they even open to receiving a callback? What happens is most users just fill out the form and then they don’t really want to be called. Filtering out those users can also really help us from a quality perspective. However, as you see there in April, 19% of the total leads that were unqualified, 19% were spam.

We wanted to go even further. That’s where the pillar four comes into play, SMS verification. For those who don’t know what this is, as Lisa mentioned, this is basically a user experience where when the user is filling out the form, they’re asked to verify their phone number before that lead is created in the CRM. Typically, if they’re bots or spam, they usually get eliminated since they aren’t able to actually input the text code that they receive on their phone.

As you can see, by June, our total unqualified leads by reason were gone. We had no spam. Our primary reason for being unqualified was just unsure, which is what we’re going to tackle next. Now, before going into the next pillar, I do want to clarify that when we make optimizations such as conditional questions or SMS verification, we should expect in-platform CPLs to increase.

The reason being is we’re going towards quality at the sacrifice of a little front-end efficiency, but that should be always compensated at the booking stage, at the new patient stage. We should be seeing lower cost per booking, lower cost per patient or new member. Once we get rid of all the spam, then basically, we want to improve the schedule rate or the new member rate, because we want to be competitive when it comes to search.

Search is always this demand capture channel, and we know social has this thing to be demand generation. We want to try to get as close to search as possible. For that, we need to look into the number one reason for unqualified leads. As I was showing you in the previous chart, now the number one reason for why leads were marked unqualified was specifically because they didn’t have insurance.

Basically, what we did here is we added insurance as a conditional question on the forms. We did it in two separate ways. By, A, showing them a full list of insurances, and if they selected, I don’t have insurance or other, then they got disqualified, or showing them the insurance that was available based on the state that they were in. If the client had insurance coverage depending on state, that’s what the user will see after they selected the state. Now, this reduced our cost per new member by 44% month over month. The next stage is continue iterating the user experience in the form, and testing different ways in which you could tackle that reason. As I was saying, we had two different types of forms that used the insurance question as a conditional question. One was showing them the entire list of insurances, and the other one was just showing them insurances based on the state that they were in.

As it turns out, and to our surprise, actually the form that was showing the entire insurance list actually had the lowest cost per new member versus the other one who had 2X the cost per new member. When you look at it from a CPL perspective, you can see that you would pause the insurance by state list, but actually the other one was the winning one.

By pausing the insurance by state form, we actually were able to reduce it 25% further, the cost per new member. This is the whole evolution of the strategy. Showing first where we started and seeing the black line go all the way down, each month trending down first 30%, 7%. The biggest, biggest change was adding the form with the conditional question in August and then continuing to optimize that and testing different ways in which we could ask that question and filter for that. Also, we’d like to note the cost. The investment level also almost doubled when comparing May to September, we’ll also have the cost per new member decrease.

The other thing, too, is we always like a healthy competition within Cardinals, so I always like to compete with search. When looking at cost per new member between paid search and paid social, when looking at the last four months, so since June to September, paid social cost per new member was actually 28% lower than paid search. Obviously, the scaling paid social is much more lower than search, but still, this is not to say that we should obviously overinvest in Meta, but it is a good component of our immediate mix strategy and can actually help our blended cost per new member go down. Passing to Lisa.

Lisa: Awesome. To tie things up, we still got a little bit of time and some Q&A as well, but overall, it’s a shift in a paradigm, really. It’s not thinking of Meta anymore or only exclusively as a brand channel. It’s seeing Meta as a viable channel to drive new patient acquisition and to drive bottom of funnel leads. It’s a channel that we can utilize additional targeting like third party in order to engage users who might not have been found on search.

There are a lot of incremental audiences that can be tapped into on the Meta side. The creative is really fun and engaging. We can use things like quizzes. We can use things like the lead forum. We can use instant experience, which is a micro landing page within Meta where you can get the users to engage with different content relating to your brand and still remain within the Meta ecosystem.

It’s really a way to just find incremental audiences with fun, innovative, creative formats. Ultimately, you’re able to then retarget users who are viewing those videos, engaging with that instant experience, and get them down to the bottom of the funnel and ultimately pan out into a conversion. A really quick slide just about some of the use cases that are not to be utilized for Meta, it’s really a discovery platform. We really want to make sure that we’re nurturing those leads appropriately.

You do still need to rely on some form of upper and middle funnel. With Meta, a lot of advertisers will say, “Let’s now just go fully into leads.” We do want to make sure that there is some form of brand education and establishing of that brand trust. You can utilize a lot of client testimonials, or that type of content which looks very organic to the platform. Then making sure that we’re priming these audiences to get them the most engaged and ready to convert ultimately at the bottom of the funnel.

We also use CRM data insights, so don’t ignore those. Make sure that you’re utilizing and looking on the backend, to see, of the leads coming through Meta, how many are becoming actual booked appointments? What is the reason if they’re not? Is it because they’re not interested? Is it because they don’t have the right insurance? Those are things that we can add to the conditional form. We also want to do testing with the form as well, including long form versus short form. There’s a lot of different optimizations that you can make. 

Rithika: Lisa, I have a few questions for you guys.

Lisa: Yes. [unintelligible 00:25:28]

Rithika: Caro, this might be for you. If people didn’t have insurance, where were they directed to next on the questions?

Caro: That’s a really good question. We don’t want to lose, because we pay by impression. We don’t want to lose those dollars. We would always want to send them to a resource center, a blog, and tell them, we are also here to help. Always make sure when you’re building the form to create that thank you page for leads who were filtered out, but also send them to your website, resources, centers, blogs, at least leveraging that impression in another way.

Lisa: You can retarget them if you want, or not retarget, but you can have different messaging for people who may not be ready to convert to explain more of the educational side of why they should consider the brand. You can still nurture leads that aren’t quite at the point of conversion.

Rithika: Did the ad sets include any video?

Caro: Yes. In some cases, yes. It really depends on the budget that we’re dealing with. We have typically always found that even animated statics always outperform video, and they even outperform animated statics. If we have to select one media type, I would always, for action, go with static. Just because in most of the cases where we had enough budget to test both, static always won.

Lisa: For conversion.

Caro: Yes, for the conversion portion. [crosstalk]

Rithika: Did you have to build totally different creatives for Meta, or were you able to repurpose what you were already running on Google and other channels?

Caro: I would say definitely try to build creatives for Meta. Specifically for the first pillar, you really want to be intentional about what you’re offering to them and prioritize the user experience. You don’t want to be serving square ads, story placements, or have the Meta interface be overlaying the main text that you have or the main message that you have. That would be my recommendation.

Rithika: Last question. It’s a two-part question. How can Meta lead forms be used in a compliant way? Is it asking people what services they’re interested in, in the Meta ad itself considered PHI?

Lisa: Meta is not collecting any of the PHI. It is being passed directly back through to your CRM data. It’s only being touched or being seen within your CRM.

Rithika: I said last questions, but we actually got two really good questions come through after that. When you were running those A/B tests on messaging, how long did you let them run before deciding what actually worked?

Caro: That’s a really good question. Typically, we’d like to do between one or two months. Again, it depends on the investment level that we have on the account. Also, if you want to make sure that the test is conclusive, you can always use statistical significance calculators to make sure that you have a clear winner.

Lisa: We use those a lot. One of the other points from what to avoid on Meta, make sure you don’t split the budget too small. You need to make sure that there’s proper funding to actually run an adequate test. There are actually online statistical significance calculators that you can utilize that will tell you at which point it becomes a statistical variable that you can actually say with certainty one over the other. Make sure that you’re not segmenting your audiences or your ad sets too small, so that you’re really giving the algorithm a fair shot at determining the winner.

Rithika: Lastly, since statics convert better, would you recommend your standard static creatives or more interactive ones like the Instant Experience ones?

Caro: The Instant Experiences would be more for a mid-funnel stage since they are like a micro landing page. Then statics would be for bottom funnel stage since it’s action-oriented.

Lisa: We usually do video in the top of the funnel, and then we target the video viewers in the middle of the funnel, serve them an Instant Experience where they can engage with the content, learn more about it because they’re showing intent, but they’re not quite ready to convert. Then at the bottom of the funnel, we serve the people who engage with the Instant Experience, a lead form.

Rithika: Perfect. With that, thank you, Lisa and Caro for an extremely insightful presentation. 

Caro: Thanks, guys.

Lisa: Bye, all.

Rithika: Bye, all.