Episode 06: Pioneering Analytics for the Privacy Era
Hosted by Aaron Burnett with Special Guest Laura Chase
Dig into the ever transforming realm of digital analytics and marketing, as privacy regulations continue to shape the way companies collect and use customer data. Laura Chase, Senior MarTech Strategist at Wheelhouse, emphasizes the importance of embracing privacy rather than fighting against it, while advocating for healthcare organizations to take control of their own data by implementing private client IDs and being selective about the data they share with third-party vendors. Throughout the conversation, Laura shares her experiences and insights from working with various organizations, stressing the need for analysts to understand the data they work with and for marketers to move away from “lazy” tactics and focus on creating genuinely relevant experiences for their audiences.
Laura Chase’s Background and Experience
Welcome to the Digital Clinic, the podcast that goes deep on critical digital marketing trends, strategies, and tactics for the healthcare and medical device industries. Each episode brings you expert guests sharing the knowledge, insights, and advice that healthcare marketers need to be successful in this complex and rapidly evolving digital landscape.
Aaron: With me today is Laura Chase, who previously was Director of Web Analytics and Insights at Providence and has joined Wheelhouse recently as Senior MarTech Strategist. Laura, thank you for being with me today.
Laura: Thank you for having me.
Aaron: I actually am really looking forward to this because you have such an amazing background, and I get to ask you questions about your background. I’ve never printed out anyone’s LinkedIn before to ask them questions, but I want to know about everything you’ve done. If you don’t mind, let’s start at the beginning-ish. Let’s go back to Adobe and what you did at Adobe and kind of march forward. In particular, I’m interested in what you learned along the way that gave you the perspective that you have now.
Laura: Yeah, so I fell into analytics. It wasn’t a thing that you studied for. There wasn’t an analytics job at the time. This was like the early 2000s, like 2001, where I was working for Backcountry. We were using Hitbox at the time, and they were just moving over to Omniture. I was doing order analysis and working for the finance department. I taught myself SQL, and the marketing team found out that I knew SQL, and they were like, “Well, we have some questions for you,” so that’s when I was like, “Okay, we’re going to use this Hitbox data, and we’re going to tie that to our database data.” We’ve got healthy SQL queries, and we’re going to do that, and so then we’re migrating over to Omniture. I found out Omniture was my backyard at that point, and I thought, “This stuff is awesome.” So, I went over to Omniture and had a great time.
You kind of asked about my perspective, like what changed how I think about analytics, and I think that’s when I was there, and I was a business analyst. Purely business analysts were running reports, were doing analysis for clients, and I just couldn’t make heads or tails of the data because there was this data, but I don’t know what it means. I don’t know how it’s being collected. I don’t know how it’s being processed. I’m just getting the output, but I don’t know exactly what that output means. So, if I’m joining columns of data, back in the day, if I’m using this data, I don’t know what exactly that means. Is it a click? Is it a load? What is that? At that time, it was still the beginning of analytics, digital analytics, they had their technical implementation, an analyst, and they were on one side, they were like the developers. Then you had your business analysts, and they didn’t really talk to each other. I think at that point, that’s where I was like, “This doesn’t make any sense to me.” I need to understand how the data works, so I know how to use the data. So, I did a deep dive into the tech side. I had a really good friend I got hired with who was on the tech side and taught me everything. It was awesome. I figured out how to do the implementation. Then I figured out, talking again with, with different consultants and with different product managers, how the data was being processed, and the light went on. I was like, “Okay, so that totally makes sense to me now,” what I’m collecting and how that spat out on the other side. Now I can make sense of this data. I was probably there a couple years, at what became Adobe, before I figured that out. I think that was probably the biggest thing that swayed how I think about data.
Now, even how many years later, and running teams, this is the first thing I’ll do with my analysts is check, “Yeah, okay. They’ve got some great SQL knowledge. They’ve got some great Python. Do they understand the data that they’re using?” Then going back or working backwards and making sure that they understand. If you’re a digital analyst, you really, really need to understand your data collection. Can you open a debugger? You’ve got great scripting skills, so if you can do SQL, if you do Python, you can do JavaScript. So, we’ll work backwards, and we look at that data collection, and then we go forward. I want to say, walking that hybrid line between tech and business, it’s opinionated. I feel like it’s made me a better analyst, but I’ve also seen my analysts that have worked for me, and once they can walk that that road in between tech and business, they become better analysts, but they also become, honestly, better storytellers. When they start to understand the business more they’re like, “This is capable. This is not capable. Oh, can we collect that piece of data that my business stakeholder needs? Well, how do I collect that?” So I’m actually answering their business question, not just throwing smoke and mirrors at it.
Aaron: What was the company you were at that was acquired by Adobe?
Laura: It was Omniture, way back in the day.
Aaron: I did not know that.
Laura: Omniture. It was based in Utah, and I was working out of Park City, and Omniture was out of Oregon. It’s just through the mountains.
Aaron: I think, well, I know one of the cofounders of Omniture is also behind Domo now.
Laura: Yes, and the other one is at ObservePoint.
Aaron: I didn’t realize the other was at ObservePoint.
Laura: Yes, John Pestana. I met him at their EDGE, kind of like their summit event. Awesome company.
Aaron: Yeah. I met him years ago as well, when I was running marketing for a company that won an award where Omniture also was winning an award. Interesting guy.
Laura: Yes.
Aaron: From there, then to Overstock, where I assume, again, it was all about answering the business question?
Laura: Yeah. Well, I did Adobe consulting about five or six years, went to Singapore with Adobe, but mostly it was working with banks in Singapore. I was in finance and banking, and I worked with all the big banks, all of them. I was working with Citibank, and I actually got to know the Long Island, New York team while I was in Singapore. I was on midnight calls with New York while I was in Singapore, so after I moved back to the States, I actually worked with Citibank. Every experience creates how you think about things, but I was working with the global team. That was really, really interesting. I think that was when I started thinking about, “Hey, there’s different digital maturity.” I’m working with all different countries, right? We had Australia, and at the time, they were doing A/B testing and optimization and trying to automate as much stuff as they could. Others were talking about building a website, what it takes to build a website, do you have an app? Have you heard of analytics? We had the whole gamut of immature to mature. How do you move organizations from knowing nothing and that education to move little by little by little by little to get into a mature organization?
Aaron: So, from there to T-Mobile?
Laura: Yep. I did Overstock as just kind of a layover. My family was tired of me traveling internationally. I love traveling internationally, but I had toddlers, so my husband at the time did not love my international travel. I worked for a year for Overstock. They’re in e-commerce, and it was kind of fun. I started in e-commerce. I was a consultant, and I worked in all the different industries, grounded in financial services. Financial services, at the time, were slow, slow, slow, behind the curve, slow adopters, conservative. You want your banks conservative, that was fine. I was okay and happy with being conservative.
Then jumping back into e-commerce, it was like, “Oh my gosh, we’re at running pace again, and now we’re going really fast.” That’s probably the time I got really into A/B testing because I was running both teams. I had the analytics team and the A/B testing team. I think that was the part where I thought what was really cool about having both of those teams was with an analytics team, it’s really easy just to become a reactive BI reporting team. The value isn’t really seen in that, right? You’re a cost center, you’re not a profit center. So it’s like, “Well, how do we change that conversation so an analytics team becomes a profit center instead of a cost center?” That was by doing some of the analysis to make recommendations to the product managers and stakeholders about, “Hey, we were seeing an area we could optimize. We want to test that.” They had a testing culture, which was great. It was already a mature testing culture, and so that’s where you can get that cycle going. Every insight should lead to a test, every test should lead to an insight.
Even at the time, the product managers could go through and it became very mature and even a part of compensation to say “Hey, I have a thing, and I’ve improved that thing.” How much of the bottom line did a particular product manager contribute? That could help with bonus structures and performance structures. That was really cool because in my head, it changed that conversation of an analytics team being a cost organization to a profit organization because we’re constantly turning out optimizations. Morale was great; the team was great because it was exciting. You’re working hand in hand with your stakeholder and with your product manager, not, “What reports can I give you,” but “How do I help make you more successful?” That’s when analytics becomes fun, when the analyst starts to have some skin in the game, the analyst starts to think, “Hey, wait a second. I want to make my stakeholders more successful. I’m going to be thinking about data. I want to be proactive.” We’re looking at the data, and yeah, we need to hand over reports, absolutely. There’s the basic scorecard. I’ve got to deliver that reactive standard data, you just need to do this. But, when I’m seeing things as an analyst, maybe the product manager isn’t seeing everything, maybe I’m seeing areas of opportunities, I’m seeing areas of fallout, I’m seeing targets or segments of the population that could be doing better or worse. If an analyst is able to bring that to the table proactively, and you’ve got a product manager who’s willing to listen, that’s when you really start to see that improvement of, “Now I’ve got my conversion improving. I’m starting to see different targeted audiences behave better. I’m starting to cater to them more specifically.” That’s when you start to see the magic happening.
Aaron: To clarify, you experienced this certainly at Overstock. Did you have a similar experience then at T-Mobile?
Laura: Yeah, we did. It was great. Again, I was running the testing team and then inherited the analytics team, so we did it a little bit backwards there, but same sort of thing. It was kind of cool because I did have a really mature testing team, and they were coming up with their own proactive ideas to test. They were very, very data driven. What is the data saying? We’ll go run the data. Yeah, we ran a lot of tests that were just like stakeholder says they want to run, “Is blue button better than green button?” If it helps them adopt testing, fine, we’ll do that. It’s not the most valuable test, but it gets their toes in the water.
Aaron: It sounds like a very Google-ish test. Testing shades of blue.
Laura: Exactly. Or, “I want to test if size nine font versus size ten font is going to make people click on this button.” Sure. It’s not going to make a difference, but absolutely. We’ll get you tested. That’s where we started to get into saying, “How do we do more personalization and optimization, not just true A/B testing?” I had a data scientist who was working for me at the time, a junior data scientist, and it was a lot of fun. That was at the beginning when Adobe, for one minute, they had a BYOA, bring your own algorithm. They were going to implement that in Target, and we were just like, “Yay!” I had this really bright guy working for me, and I asked, “Do you think we could be part of the beta?” and he goes, “I think we can just do it.” I had a data scientist that was working for me, just a junior data scientist, and I’m like, “Alright, you guys go figure this out. See if we can do this.” It was awesome that we did have that latitude. My director that I was working for at the time was like, “Yeah, let’s go see if we can do this. I want to see you be innovative.”
I think it takes the right kind of leader to give that bandwidth to say, you’re going to go outside of stakeholder request only, to we’re going to see if we can invent something new and if the payoff is there. I’m very grateful for having that director at that time because he allowed us that. It was great. The guys came back, and I was like, “Go through all the attributes that we’re collecting on the website and analytics. Go through all of those and find those that have the greatest propensity, to actually determining what behavior on the website somebody’s doing. We wanted to see if somebody signed up for a plan or are they buying a new phone? There are these major journeys. He went through, and he came back, and I was expecting we were going to have like 25 or 30 of these. He said there are six. Six attributes that are really determining the major journey. We were chasing a lot of tail, we were like “Hey, this is proof of concept, let’s see if we can do it.” So, we had a quadrant – we had maturity, a new customer, a repeat customer, or somebody buying a new plan – we were able to put all of the customers into a quadrant. He created his model, and we’re like, let’s test if we could run this. The other guy who was kind of a developer of computer science was a really, really smart guy. They created some JavaScript, we put it in through Target, and it worked. It was awesome, but that was the first foray of “Let’s get scrappy,” and see if we can optimize in real time on the fly by just doing it.
Aaron: By identifying these six characteristics that would indicate that people are headed to one or another very important user journey, you were able to assign them to one of four quadrants.
Laura: Yeah, it was super basic.
Aaron: Then to tailor, I assume, their user journey based upon that initial assignment.
Laura: Yep, so then you could say, “Hey, based on where you’ve come from based on these six different attributes, we’ve got different banners, and we think you’re most likely to be interested in…”
Aaron: Which sounds very much like the training wheels version of CDP today.
Laura: Oh, yeah. It was totally training wheels, maybe tricycle?
Aaron: Well, I think the challenge there is not just business and data, it’s having the ability to think in a very rigid and systematic fashion and overlay that with creativity, coloring outside the lines.
Laura: That’s a good way to phrase it. I like that. To blend that, I feel like that’s where the power comes from. You’re not just building a thing, you’re not just marketing a thing, but you’re able to blend the two to say let’s build a thing people actually want.
Aaron: So, from there to AWS. Tell me about the role there and what you learned.
Laura: I learned about transformation. There was a rule, I want to say I read it somewhere like Quora or LinkedIn Learning. It’s not how fast an organization should transform but how fast they can transform. So, keeping that idea in mind, I feel like everywhere I go, they’re like, “Hey, Laura, you’re a data person. I want you to come in and build a team from scratch.” We’re not data driven, but we want you to build an analytics team from scratch, A/B testing from scratch, and we’re going to change our whole culture. You realize the easy part is the tech. That’s the easy part. Hiring the people, the change management, the culture shift, that’s where the real work comes in. Saying, how fast can certain orgs go? I don’t have the answer for that. It’s like how I mentioned the people that ask, “You want to test green versus blue?” Absolutely because it got him in that mindset. But how much of that do we do, even though we know there’s no performance value of that activity? However, there might be cultural value to that activity; I don’t have the answer. I’m kind of noodling over that.
How do you get an organization to start shifting to get into that mindset? Overstock was very, very mature at it, but most other organizations, the concept of using data as an input to form your ideas, rather than as a reactive way to reinforce your ideas was foreign. How do you get from reactive, “Oh, this is confirming that my opinion was correct,” to, “This is informing my opinion.” Then what are the stages to go from A to B, and what does it take to get there? It’s not going to happen overnight. Some orgs might be able to do it in six months, other orgs, it might take them 15 years.
Impact of Privacy Regulations on Analytics and Marketing
Laura: That’s where I got the EU cookie policy though. So Amazon of course, they have target on their back, like any big company, they’ve got a big target on their back. EU affirmative cookie consent came down, we had GDPR and CCPA hitting us all at once. It was painful; that was super, super painful. At the time, there was me and my good friend Kalyani. We had been working together for 17 years now. At the time, we were just like, “I guess we go find new careers.” EU cookie consent, it’s all opt in, nobody’s opting in, and we’re just like, I guess this is a time for a career change. We thought about it and we thought about it, and we’re like, “Alright, we’re going to do our best.” I think that’s where, for both of us, we changed our mindset of instead of resisting privacy, what if we just embraced it? What if we said, “I’m not even going to take the basic recommendations. If I’m in, I’m 100% in.” How do we start rethinking analytics and MarTech with the consumer first. If you think about it, analytics was late 90s, and there were no privacy regulations. It was the wild west. They tracked everything and nobody cleared cookies. We didn’t know what cookies were. I remember at the beginning tracking email addresses and phone numbers.
Aaron: Which was, by the way, quite exciting from a marketing perspective.
Laura: Right, I’m horrified now.
Aaron: It seemed like a good idea at the time.
Laura: It was a great way to get all of this information, passing via public network calls. That’s where we started, and so of course, the easiest thing to do was give everybody an ID, and then we’re going to throw them all in a database. Then we’re going to sum them all up, we’re going to aggregate them, and then we’re going to spit out a number. So, how many years later, we’re like, wait a second, privacy first. Not everybody needs an individual ID. Do you really need to get the most granular to aggregate those to get the answer that you want from the top? For most marketing campaigns, I don’t need to know if Joe, Steve, or John purchased. I need to know, what was my return on investment overall? So that’s where we said, “Can we get into this idea of measuring at a cohort level?”
I went back to T-Mobile. I had my data scientist run how many attributes are really determining the journey of a customer? Every data set is a little bit different. When it came back with six, I was like, we don’t need 200 attributes. We don’t need everything about an individual. If six attributes are the main determiners, what if we just had the six combinations of cohorts, where we can start to say, at an aggregate, before somebody’s told me who they are, how I can still provide something relevant without being creepy, without asking them too much information, so I can keep that trust with the customer? There is that division of relevant and then it can slide over to creepy. How do we stay on the relevant side?
Providence had called me while I was at AWS and was like, “Hey, we are having some data problems. Could you come fix them for us? At that time, I was all serious about privacy. I was like, I love this privacy journey. This is exciting. This is fun, but you have to reinvent it. I can’t be thinking about analytics the way I’ve always thought about analytics. I can’t be thinking about MarTech the way we’ve always thought about it – I just put a cookie on and send the information. I liked the challenge. I said, “Absolutely, healthcare. Let’s do this,” and so we did.
Aaron: At Amazon, given the GDPR implications, were you able to shift instrumentation, shift the approach to analytics and data collection in the way that you envisioned?
Laura: No, no, we didn’t do it. That was another lesson learned. I think that it did help them form being flexible, trying to do your front-end implementation so that you are vendor agnostic, and so you can be as flexible as possible because on the front-end, we can move very quickly. We were like, “Yes, we’re going to change these IDs. We can implement it this way.” We knew what was happening in EU, but the back-end was so entrenched, and how they had implemented the databases and the data warehouses, they couldn’t shift on a dime. It wasn’t even an option.
Aaron: Sure. You could change collection, but you couldn’t change the infrastructure that was the repository for the data.
Laura: Exactly, especially when you’re thinking about something as important as a client ID. I think that really formed once I came to Providence and even here, what we lost is working with the private ID. That’s why it’s so so so important to, as a company, that own your own client ID. You need to feed that into your vendors. If you’re trying to do that always on the back-end, and now you’ve got lots of different IDs, and you can’t move quickly because every time somebody introduces a new tool, or a vendor doesn’t support what we want to support anymore, or now we’ve got to adopt GDPR, we can’t do that method anymore. Now you’ve got to reinvent your entire data stack, so that’s where I’m like no, if you can own your own client ID – hopefully you own your own back-end ID, too, that was one we were working on with clients, they’re using vendor IDs as primary keys and I’m like, no, no, no, don’t do it – but that way they can they can move more quickly. I honestly think that’s the primary thing that I would say clients need to do is own that and then push that ID, their own client ID, into vendors. If we’re using Google, and I’m like, “I don’t want to use Google anymore. My business use case has changed. I want to switch over to Adobe,” I can do that very easily. I’m not like, “Great. Now I’ve got to redo my entire data stack. My whole ID graph behind the scenes doesn’t even work anymore. Now I’ve got to add new columns. Now I’ve got to add new unique IDs.” You don’t have to. You’re already good to go.
Healthcare Analytics and the Privacy Era
Aaron: Yeah. We started working together when you joined Providence as Senior Director of Web Analytics and Insights. You’re mentioning this notion of a private client ID, which was the genesis for the HIPAA-compliant data solution that we jointly developed. Talk a little bit more about that. Talk about how conventionally or historically analytics and tracking worked, versus the notion of having a private client ID.
Laura: Honestly, one of my big soapboxes is about this. I know healthcare is kind of last to digital analytics. In this case, I think healthcare is leading.
Aaron: Yes, I agree.
Laura: I think everybody should implement this way moving forward, period. It should not be optional. So ordinarily, if you’re talking basic, let’s say web analytics, any client-side MarTech, you’re running a JavaScript library on your website. The JavaScript library belongs to Adobe, or Google, or belongs to Mixpanel, or belongs to SiteSpect, whatever you’re running. You’re running a client-side JavaScript library, which is setting an ID and a cookie to understand, “This is Laura Chase.” They’re running that and then collecting data that’s tied to that ID throughout every single thing I’m doing on the website – different page loads, different clicks. As I’m navigating, they’re collecting that data and tying it to that visitor ID.
Now Adobe, or Google, or whoever you’re using, will ingest all of that data into their servers, they’re going to process that and spit that back out to reports in their tool that show, “How many visits did I get this week? How many visitors clicked on this one thing, and what did these paths look like?” They can aggregate all that, and I think that’s fine. That worked for many, many years. When you come to cookie compliance, or you come to HIPAA compliance, now, you’ve got somebody’s third-party JavaScript sitting on your site, which means they’re creating their visitor ID and sending that information which includes IP address, the full URL, anything about this person, they’re sending that to their servers to process. That becomes a problem because HIPAA says IP address, certain URLs, any information that could be like an account ID and the open text field, all that will be sent to a third-party server, which could contain ePHI, electronic PHI. That’s part of the problem. With healthcare, that’s where we said, we can implement server side, where instead of pushing this data to, say a Google server, we can at least collect from the client into our servers, and then we can push a subset or a cleanse diversion into Google or Adobe.
But, I didn’t necessarily like that idea of if I have their client-side JavaScript, their ID service on my website, it’s still communicating to their server, which means I’m still sharing an IP address. Also, I’m locked into that vendor. If I’m using somebody else’s client-side ID, I’m stuck with them, so there were two parts of it. One is trying to be vendor agnostic, and the second part was saying I don’t want somebody else’s code on my client-side website. The relationship is between the customer and me, not the customer and Google. It’s my trust that gets burned if something happens, and Google does it, but it’s my trust with the customer that gets burned, so let’s just remove them completely. That was kind of the idea of we’ve got to have our own ID. Wheelhouse took it one step further and built the ID graph which is even better, so then we could control our own ID and then push everything server side. Now server side, that’s where we could say, I want these five variables to go to Adobe, but I don’t want all the variables. I only need five to answer my business use cases. Maybe I’ll send ten over to my internal data warehouse because maybe I have a CDP that does have private cloud. That was kind of the idea there.
I would love to take private ID even a step further, where it becomes not just an ID that’s run through your tag manager, but it becomes a client-side ID that you load with the page, so it becomes an ID service. Now, you can push that into your A/B testing tools, you push that into your personalization tool, or you push that through your Tag Manager into the tags that are running there. That would give you a native integration between all of your client-side datasets, so now you don’t have to do a bunch of ID lookups on your back-end. I can also start to push data at an ID level of, “I want to see how are my different personalization segments? I began using a different tool, and I want to push that into my Adobe dataset.” I can slice and dice my data for my personalization cohorts, and now I can do that through an anonymous client ID that I control by pushing to all of my third-party vendors.
Aaron: There are a couple of important implications to what you’re describing as well. The first is that one of the risks, one of the things that’s problematic about using a third-party ID, is that the PHI that they collect may not be directly collected from your site. It may be PHI that is developed based upon what they learned from you plus what they already knew. If you don’t have a BAA in place with them, which you’re not going to have with most major platform providers, then you’ve created a problem. The second, and you’ve touched on this a few times, is that because the ID is a third-party ID, you do have vendor lock in. You could switch, for example, analytics providers, but then all of the historical data is based on a different ID structure or ID graph, and integrating that historical data with new data is a monumental task. Whereas, having a private ID means that you have identity continuity, regardless of the vendors with whom you interact. Because that ID is yours, it can’t be joined with something that another third-party knows.
Laura: Exactly, yeah. I mean, you bring up a good point that I hadn’t even thought of. It is keeping that relationship between you and the customer. It doesn’t leave. Even if a third-party grabbed this ID, it means nothing to them. They’re like I can’t tie it anywhere else.
Aaron: Yeah, and so what do you think are the implications now for the rapid pace of privacy regulation, rollout, and development for Google declaring, finally, at long last, that they will deprecate support for third-party cookies? For everything that’s happening in healthcare, where do you think we’re headed in terms of analytics data strategy and data collection?
Laura: It’s awesome. It is about time. Wheelhouse is ahead of the game here. I think for a long time, if people were relying on third-party data, it was kind of like looking into a mirror that’s more of a portrait. Mobile hasn’t been using third-party data for a long, long time, and Safari hasn’t, so it hasn’t really been working anyway. Even though we were relying on it, it was kind of a half-truth. I know there’s a lot of resistance, and I’m not an AdTech person. I’ll state that I am biased. I do think it’s about time that consumers take back their privacy from the internet. I think there’s certain times where it does depend, and I was thinking about it earlier. If I’m shopping for a sofa, and I’m looking at sofas, for the next three weeks, I’m targeted for sofas in all different ways – on different social media sites I’m targeted for sofas, my partner is targeted for sofas, my children are targeted for sofas, everybody’s on board to get new sofa – that’s okay.
Now, let’s say I’m looking up a very private mental health diagnosis. I’m not so happy if that’s being shared with my partner or my social media, and now I feel like that’s a little bit creepy. I think for a long time, advertising did go down that road, and it was creepy. They kind of got away with it for a while, so I’m like, it’s about time. I had my epiphany AWS when I was like, “Do I quit my career?” No, we embrace this. There’s a better way out there; we can do it better. We don’t have to rely on the technology that was created in 1997. It’s about time that we do it better. I think, innovation, it’s there for us, and not just AI. I think AI can still be a little bit creepy. When they’re trying to use AI for, “I’m going to try to determine who you are without you telling me,” that’s still a little too far. Let’s respect the boundaries. As soon as I log in, I want that company to know me and they better tell me my preferences. That’s what the login is for, but until we’ve done that, until we’ve crossed that, I think just trying to not get that breach of trust. You can build your customers. How awesome would it be if at some point in the future instead of a, “Do you accept all of our cookies? Do you opt in or out?” what if there was a banner that says, “We don’t make you opt in or opt out because we don’t collect any data that would require that.” What would that do to your competitors? If you’re like, “Yeah, I don’t collect any private data, period.” I’m not going to get rid of analytics or advertising or marketing. I still want to create a very relevant experience without crossing that line. How awesome to shame your competitors? We’re not collecting any private data, so you don’t have to opt in or opt out. You just don’t do it.
Aaron: I think that’s interesting to explore. Thinking about healthcare specifically, it’s not the act of collection that’s problematic. It’s the act of sharing that’s problematic. You could create a paradigm in which you’re collecting everything that you need to understand the use and utility and experiences on your website, but also just have a declared policy that you don’t share that data with anybody outside of your four walls.
Laura: Yeah, exactly. I think we don’t want to do the, “Hey, we’re not going to collect any data.” I think the expectation from customers is I want a relevant experience. I don’t want something that’s completely like, “I can’t find anything I want.” You’re targeting me for stuff that doesn’t even make sense. So, still creating that relevant experience, but not going into that line of, “I didn’t even tell you that. How do you know that about me?”
Aaron: This year, let’s say things play out the way that they’re supposed to, the way that they’ve been declared or announced, which they won’t because they never do, particularly where Google is involved, but Google deprecates third-party cookies, the Privacy Sandbox is implemented, things continue as they have with analytics and healthcare, which they need to given the OCR guidance and HIPAA implication to patients and non-patients. If we were to project two or three years into the future, what do you think data collection analytics looks like? How is it accomplished? What do we have? What do we not have? How are we effective?
Laura: There’s the part where, with technology, I think Dustin and I could solve it, but that doesn’t solve the change management. Healthcare and culture are still going to take a little while. The way I would like to see the industry go is different websites and different companies start taking back all of their customers’ data from Google. Google’s looking at their flock IDs and stuff like that, which is basically cohort IDs, but they’re doing it in the browser. They still control the game, right? I would love to see if companies are doing more, like healthcare, that says, “I’m collecting enough to be relevant. The trust is built between me and my customer.” You’re saying, I’m not sharing that data. That data, that’s our relationship. I’ll keep that data internally. I’m not going to share that data with Google, and I’m not sharing it with Facebook. I’ll swap some random IDs with those, but your data is safe with us. I think healthcare is starting it because they have to, but I would love to see all websites ending up going that route.
Aaron: We’ve certainly been pushed, not just in that direction, but to do just that with some of our larger healthcare clients, with regard to advertising because we have access, we have a HIPAA-compliant data warehouse, and we have access to PHI. We can learn things from that data. We can learn things specifically with regard to the efficacy of the advertising we run, the channels in which we’re performing well, the messages, and also some of the targeting that we’re able to do, but we have a wall between analytics and advertising. Our analysts can learn, and they can then apply what they learned, but they do it without the data. They do it through synthesizing and then making decisions. That’s worked really well. It’s learning a different way to go about this and also creating a context in which they have access to data that they can easily synthesize and interpret and then turn into action. They don’t have to roll around in spreadsheets like they used to years ago. It still can be fun, but not when the spreadsheets get too big, which they are today.
Laura: No, no, I do like that because if you’re using the data for analysis, that’s fine. There’s a rich amount of just healthcare data in general. That relationship is between the caregiver and the customer or the caregiver and the patient. That has nothing to do with Google, it has nothing to do with Facebook , and it has nothing to do with advertising on ESPN. That’s that finite relationship. I think safeguarding that relationship will be paramount.
Aaron: Have you looked too much at the Privacy Sandbox?
Laura: I haven’t recently. I knew that they were working on Flok a couple years ago, when I was like, “That’s my idea, but I just didn’t want to do it at the browser level because they’re kind of solving it but they still control the data.”
Aaron: That’s exactly what they’ve done. The Privacy Sandbox is still, “Google has all of the information then you get none or very little.” It’s still, to a significant extent, this unconventional ad ecosystem, but instead of being server to server, they moved it all into the browser, so it just lives in Chrome.
Laura: Okay, yeah, yeah, that’s it. I still don’t want Google to have that data. It goes back to, for me anyway, and I would hope especially healthcare since it is such a sensitive topic, is that the relationship doesn’t belong with me and Google. The relationship belongs with me and my caregiver, me and my doctor, so it’s like keep those third parties out.
Embracing Data Privacy
Aaron: There’s so much about this, that feels to me, like a return to more conventional interpersonal relationships, where you actually strike a relationship with a provider or an entity. You give them permission to communicate with you, and then from a marketing perspective, a return to the fundamentals of good marketing, which is you understand your audience, you are very thoughtful about the messages that you develop and the creative that you use to convey those messages. You pay very close attention to that, and you tune based on creative. As opposed to, in digital advertising, we have been for far too long, able to be quite lazy. You just come up with a campaign as quickly as you can, and five variations of that campaign, and you throw it out there at something that approximates the audience that you want to hit. Then you look at the data and you refine based on that, kind of a spray and pray method. You won’t be able to do that anymore, so you have to be a better marketer, which I think is good for everybody.
Laura: We’ll have to up our game.
Aaron: That’s right.
Laura: It’s true. I love that you call that lazy marketing because I would do the same thing. I have two teenagers. We’re looking at social media and then all of a sudden my partner, a 52-year-old man is like, “How come I’m getting teenage skin care recommendations?” Oh, that’s for the 14-year-old upstairs, and I get so angry because that’s lazy marketing. You’re just using IP address, that’s lazy. You’re on social media, you already told them everything you’re interested in, and they’re still doing it. I love that you call it lazy marketing.
Aaron: Yeah, exactly. So, as we think about folks who are in analytics and healthcare, thinking about how to navigate the current circumstance, thinking about the steps to put in place now, the things to put in place now, to anticipate what will happen over the next few years. What advice would you have for them?
Laura: I would say embrace the privacy. Don’t fight against it, embrace it. There’s a way to be a better analyst. There’s a way to be a better marketer. We don’t have to hang on to the old things. Also, I think the next goal I would recommend is controlling your own data. For so long, I think we relied on, “Oh, I’ve got two lines of JavaScript,” and we had to use whatever the vendors told us was available. I think now there is the option of saying they’re a tool. They’re a good tool to get to the end, but I’m not going to give them my data. Some of them are really good at data collection, some of them really good at operations within the data, but ultimately, it’s the company’s data. We don’t have to share it. We can use them for what our needs are, giving away the baby with the bathwater.
Aaron: Any particular direction or recommendations from a technical perspective? Ways to frame technical decisions that they’re making, MarTech decisions that they may be considering?
Laura: I would say, definitely own on your own visitor ID, period. That’s number one. That’s absolutely paramount. I think from a technology standpoint, it’s like what direction do we do? Trying to get in a data system, whether that’s a Tag Manager, it doesn’t matter what it is, where, you’re controlling your own data. You’re able to pare down, “Which data am I sending to which vendor? Having that control, again, it’s the control of your own data. I’ve got some vendors where I just want to send a light amount of data. I’ve got other vendors where maybe I need more of that data. That way, I’m kind of controlling what data goes where, but all that costs a lot of money. It goes back to the cost center. How do you, who’s usually seen as a cost center, start to break into those conversations? Your leadership team saying, we want to be seen as an opportunity center. Absolutely. 100% start today, start tying your data and your analytics to testing so you can measure the value of those insights. As soon as you can measure the value of those insights, now you’ve got a dollar amount that you can go back to your leadership team with that says this is the value of this program we implemented. It’s not just sunk costs that we kind of have to right every year, we’re actually seeing profits come out of this previous cost center, so we can keep the machine running.
Aaron: Right. That makes a lot of sense. It’s good advice. Well, thank you for talking with me today.
Laura: Thanks for having me. This has been a lot of fun. I get to talk about things I really like.
Aaron: Yeah, exactly!
Questions or Comments
Please let us know if you have questions or comments about this episode by emailing Grace Johnson at grace@wheelhousedmg.com. Want to be a guest on a future episode? Fill out the Be a Guest form at the top of the Digital Clinic page to submit your inquiry.
Make sure to subscribe to the Digital Clinic on Spotify, Apple Podcasts, or Amazon Music to stay up to date on our weekly episodes.