Why Good Measurement Starts Outside the Platform

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.

Lauren Leone: Hey, everybody. Welcome back to Ignite Healthcare Marketing Podcast. I have a frequent guest, a face you guys have seen many times, our Director of Analytics, Alex Kemp. Alex, thank you for joining us yet again to dig one click deeper into some of the analytics problems we see today.

Alex Kemp: Yes, happy to be here. Thanks for having me.

Lauren: This is a fun one because we’re going to talk about some of the traps, myths, and misconceptions that many of our clients, prospective clients, industry friends tend to fall into, and look for some solutions to some of these common problems that marketers need to watch out for. There’s multiple layers here. There’s the challenges with getting the right data in the platforms. There’s the challenges with the platform tells me X, the business tells me Y. There’s the challenge with scaling your channel mix and the walled gardens, and how do I understand across channels what’s happening?

There’s an overwhelming amount of options on how you can measure, and sometimes that creates analysis paralysis for people. I want to spend some time today just really picking three very specific topics and getting your opinion on those. The first one is a fun one, and I think it’s one we’ve been talking about for a while. I’ll call it trap number one, which is the blended CPA trap. I think that’s a broad topic to start at.

The general framing of the statement here is I want to scale my media. My CPA and platform looks good. I want to add more in. I’m going to get the same CPA, right, Alex? I’m going to toss more money on it, and it’s just going to scale in a linear fashion. I think that tends to be a misconception for people when they think about sticking with a single platform like Google Ads. There’s the frame. Tell me a little bit about what you see the problem statement being, and then we can talk about some of the solutions.

Alex: Really, the problem statement is what you referred to there. Not thinking about each incremental patient that you’re getting with your media. You’re just thinking about aggregate performance month over month or week over week. You see an increase in costs and an increase in new patients, but you’re not thinking about how much did each one of those, in theory, cost with that incremental spin I had.

It’s very easy to get in the trap of, let’s say you have a $300 cost per new patient goal, and you’re at $200 feeling great, and you just keep on pushing it, when in reality, depending on the volume of your ad spin and your new patient volume, you very well could be actually losing money on each one of those incremental new patients. That is just looking at it within one platform as a silo. Then the other thing is that in conjunction with other channel that you’re running, and you’re not thinking about marginal cost per acquisition, these other channels, meanwhile, could have a more favorable marginal CPA, and that leads to the wrong distribution of budget, in theory, about not as optimized.

Lauren: Yes. It’s easy to pull up Google Ads to look at that spend and CPA trend line that they pop right in the top of your campaign table and say, “Okay, I got spent this much, and my CPA went up a little bit, but it’s still within my target.” For a marketer who’s not familiar with marginal cost per acquisition and that metric and understanding that metric, where should they start? What are some of the fundamental reports they can pull or data they can extract to start to understand that?

Alex: Sometimes it can sound like a scary term or like it’s super complicated, but really, it can be boiled down to a month over. Let’s just take a very simple example. You spent $10,000 more this month on your media, and you got X more new patients or X more qualified leads or whatever that is. It can really be simplified down to just the math of what was your increase in cost and what was your increase in new patients, and look at it that way. That’s a very simple way and a good starting point. Instead of just looking at a table of by month, here’s the performance, you also have a column for a delta. Here’s the increasing cost, increase in new patients, and then the subsequent incremental cost.

Lauren: When you see your spend went up, let’s say, 10%. We added that $10,000 on $100K budget, but your volume of patients, let’s say you only got 10 new patients for that $10,000. Blended, it’s still going to look pretty favorable. It’s going to be sub $300, but your marginal cost was $1,000, 10 patients, $10,000. Then you want to start thinking about, could I have spent that money elsewhere? Maybe my cost per acquisition on Meta is $500, and I don’t love that compared to my $200 on Google or my $300 cap, but in theory, spending it there and then blending that back with my Google ads could have yielded a lower blended cost per acquisition for my channel mix as a whole.

Alex: Exactly. They could be continuing to push into search versus looking at these other channels in a different perspective with marginal or incremental cost per acquisition.

Lauren: You laid it out pretty simply. Look at your month-over-month, these two metrics together, and do the ratio. Any other warning signs and platform that people could look out for, things like impression share and other metrics that might be indicative that marginally they’re inefficient, even if the blended looks good?

Alex: Yes. I would say your impression share loss to budget, that is advantageous if you have any of that in your Google Ads account for your search campaigns. Google is essentially saying we can pretty much give you the same efficiency if you just give us more budget. When that gets to zero, and you will have 60% loss to rank, and you have to start getting into loss to rank in order to spend more and get more new patients, that’s definitely a warning sign, I would say. That’s like the canary in the coal mine of higher CPCs, lower conversion rates, even higher CPAs than what you’re currently getting on your incremental. Yes, I would definitely say impression share loss to rank is something to look out for.

Lauren: Yes. I’ll give the caveat that I always do that, if you have a high impression share loss to rank, it doesn’t mean I’m done with this channel, I’ve maxed it out. It means start digging into why your rank is a problem-

Alex: Exactly.

Lauren: -and spend some time on the things that are just throwing more money at it, like your copy, your quality scores. Your keyword relevance. All those other fun things that the platform might not necessarily tell you. You think you have this great optimization score, but there’s all these other things you could dig into. Call out for anybody who is seeing that high impression share loss to rank, it doesn’t mean it’s a loss cause. That’s really good advice. The problem I hear all the time is, “We added $40K to our budget last month, and we didn’t get anything for it. We just have nowhere else to go.”

We’re getting the most we can, and that’s not necessarily true, but it does tell you what next steps you should be taking. Speaking of platform metrics and the gotchas there, trap number two is the overstatement when you look at your platform. Let’s say you’re running Google, Meta, and maybe some programmatic, and you go in, and at the end of the month, you’re pulling conversions from all three platforms. You pull them into your report, and you blend it with your spend. We got 2,000 conversions. You go to your Ops team or your CFO, and they’re like, “We only had 1,500 new patients.” That’s literally impossible. How could that happen, Alex?

Alex: Each ad platform has their own incentives for various reasons to take as much credit as they can, as it pertains to an ad exposure, and then someone doing something that we want them to do. For that reason, and this really excuse more, I would say, on Meta and programmatic side, where view-through conversions is a big part of how Meta and other DSPs like to measure within their own platforms.

Lauren: As they should, but it just is a trap. We don’t know how it works.

Alex: It’s valuable information. It’s just not what you want to use to pull how many new patients you got across your media. That is definitely a trap. Again, Google can do this as well at times, but to a lesser degree.

Classic example is if someone searched a non-branded query and converted on that session, Google Ads is going to take credit for that. The other thing that Google Ads doesn’t know and the other Meta and the DSPs don’t know is that person was served impressions two, three, four, five days ago. They’re measuring. They said, okay, that person, even though it wasn’t on our click, they did the thing that we wanted them to, even after our impression.

Lauren: After seeing them.

Alex: Then you get one lead or one new patient, but three platforms saying, I have one.

Lauren: Three platforms saying each one which would be three if you just assumed that the truth coming out of every platform was an individual new patient. There’s a lot markers can do in terms of the settings, the attribution windows. Any tips on setting that up to try to narrow that, but also to not starve each individual platform from the metrics it needs to run and grow?

Alex: Yes. Setting your look back into your view-through and click-through attribution windows to something that is realistic for your business. I don’t think there’s a one-size-fits-all, but usually we stick to a one-day view, a seven-day click. I believe that’s what we go with. I think it depends on whatever KPI you’re looking at and the cycle.

Lauren: If it’s a long multi-touch journey, you might have a larger window.

Alex: The other thing I would point out, too, is in Meta, you can look at click-through conversions versus view-through conversions. In their event, it’ll just count it as all one bucket, but you can’t split that out to start seeing, okay, how many click-through conversions I’m actually getting from Facebook. It’s still not the source of truth. It just provides more context about what’s view-through versus click-through.

Lauren: Pulling that into your dashboard or whatever you’re looking at might help you at least close the gap on why your number doesn’t match the business’s number. We like to create an assist column. Calling those view-throughs assists. They have a value. You can potentially even assign a value if you’re doing some sort of data-driven model, but to recognize and acknowledge that channel’s contribution without saying, I’m going to use this to say it entirely and solely created this lead for the business.

Final thoughts on this topic. A way to get around this is obviously to use offline conversion data as much as possible. We know that’s tricky here, but what would an ideal setup be to have platform metrics working for you, but also, if you want to really get down to how much we’re driving and the business metrics? That’ll take us into our trap and myth number three here, which is, platform metrics don’t always map to business outcomes. Maybe we jump right into that because I think that is the solution to number two, which we just discussed. What should businesses be thinking about in terms of using and getting as close to business metrics as possible?

Alex: Yes. I think there’s different ways you go about this. The ideal state is really what you’re looking for is a single source of truth, usually going to be something-

Lauren: CRM.

Alex: -like a CRM.

Lauren: Ideally.

Alex: Basically, housing every lead and the stage of that lead, whether they convert it or not. Then the other harder part is the technical details about UTM parameters and click IDs that are making their way into that CRM. What that accomplishes is you’ve established a single source of truth that the actual leadership team is using to define success and measure media. You’ve established channel attribution within that source of truth. Then, if you use that source of truth to go into your point about making those metrics work for you in the platform, you can start using that as your source of truth for pushing data back into platforms.

It’s already following the established channel attribution methodology. That way, you’re not using pixels or specific platform mechanisms to measure something. You’re using the single source of truth that is outside of all those platforms. That way, you’re not overstating Google Ads or understating other platforms. You’re giving it as much credit as the source of truth is, as well.

Lauren: For those of you who may be new to listening, we’ve posted many episodes on how to get that data into platform in a HIPAA-compliant way, so we’re not suggesting just uploading PII or PHI back into platform. That’s a whole other topic, but lots of ideas and thoughts there for anyone who’s listening who wants to know more on that. The single source of truth is an ideal end state, but it still is going to rely on some mostly last-click attribution. It’s going to look for that. When you are looking to build a multi-channel approach, that approach is never going to give credit to Meta for all those view-through conversions that it drove and the way that it assisted the journey.

There’s some things that we’re doing a lot more of in terms of layering on top of. It’s an and statement, not an or statement for measurement. Can you talk about some of the media mix modeling and incrementality testing that really should go hand-in-hand with that source of truth so that you can get your C-suite and your board information on what you really drove, but that you don’t rely entirely on only on that to help you drive your mix?

Alex: Absolutely. That’s an important piece because even as good as everything I described, as good as that would be, it’s still not the whole puzzle because, to your point, it will understate platforms like Meta and DSPs because of just the inherent nature of those platforms and how impressions are much more valuable for those types of platforms for speed search. Impressions are not being considered in a CRM. That’s a big limitation.

We do think of it as a triangulation mindset or concept of attribution, in this case, last click is just one piece of the puzzle. It does tell us something, but it doesn’t tell us everything we need to know. Media mix modeling is where we can really start. That was the other way I was going to say. To look at true or modeled channel contribution for revenue or for new patients is going to be more context than what you have with attribution at the end.

Lauren: One doesn’t stand alone as the perfect solution. It’s when you have them running concurrently and understand the relationship between the money you put in and the revenue that you drive for the business, that correlation, plus the last click attribution, to really get down to who exactly is coming in the door and how much are we spending to get them there.

Alex: Yes, exactly. Then the last piece of the triangle being incrementality and incrementality testing. Actually validating that MMM model and calibrating it as well. Actually doing geo-lift tests or hole-out tests to understand what exactly is this channel contributing incrementally on top of a baseline. That’s the other piece of the puzzle that completes the measurement triangle, I think.

Lauren: Yes. Again, topics that we spend entire podcast and webinar episodes going really deep into. If that’s of interest to you, check out our library because you’re going to find topics specific to those. Alex, before I close this out here, I skipped a little bit ahead because that triangulation gets me really excited. When we talk about the platform metrics and the challenges there, the over-reliance on the lead, a click, a form submission, a call, having businesses move closer to that business metric. Even if it’s a call tracking solution with AI listening capabilities to tell you if the appointment was booked or if the patient was new, tracking and tagging your online scheduler so that you can understand sourcing there.

Of course, CRM is like the holy grail because then you can follow the patient through whatever that initial lead is. It’s good to have a healthy amount of skepticism of the data coming out of platforms. It’s good to always ask, is this the best way to measure? Is this the only way to measure? Is there another layer I could add here to grow my trust and my organization’s trust in what marketing is driving for the business? I encourage everyone to, if any of these myths or traps apply to you, to dig a little bit deeper and see if there’s a solution, a way you can advance or add on to your measurement framework.

As we like to call it, every marketer should be asking these questions on a regular basis. I think some of our favorite partnerships and relationships are where this is a central topic, and we’re talking about measurement framework all the time. We all want to understand and get to the right source of truth. It was great to have you and appreciate your time. If you guys are listening, subscribe, follow us wherever you get your podcasts, and we’ll see you next time on Ignite.

Alex: Thank you.

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.