What is First Party Data Second Party Data and Third Party Data 21

{PODCAST} In-Ear Insights: Marketing Data Violating Privacy Regulations

In this episode of In-Ear Insights, Katie and Chris discuss the trend of marketing data consolidation, customer data platforms, and how marketers may be unknowingly violating multiple privacy regulations such as GDPR, CCPA, HIPAA, and COPPA in their quest for “one view of the customer”. Learn more about some of the regulations, what the penalties are for violations (including jail time), and how to protect yourself and your company.

Noteworthy links:

[podcastsponsor]

Listen to the audio here:

Download the MP3 audio here.

Machine-Generated Transcript

What follows is an AI-generated transcript. The transcript may contain errors and is not a substitute for listening to the episode.

Christopher Penn
In this week’s in your insights, customer data is all the rage, how can we get more of it? How can we bring more of it into our systems? How can we get a Master View of the customer and collect as much data to train AI and machine learning on? How can we even do things like influence elections by gather gay people to install silly apps on their phone, send us photographs of themselves, and, and turn that into actionable intelligence. All of this sounds great. But it sounds like the potential to go really, really, really wrong. So Katie, talk us through this. What are the things that people are just not thinking about when it comes to customer data and this relentless pursuit of gathering as much information on people as possible?

Katie Robbert
Well, so a little bit of background, for those that don’t know, I actually spent the first half of my career well over a decade in the pharmaceutical and medical and regulatory space. And so I know lot about protected health information, personally identifying information, and then HIPAA, HIPAA stands for the Health Insurance Portability and Accountability Act. And you know, so if you’ve ever been to the doctor, you typically have to sign some sort of a form saying that, you know, you understand your rights with HIPAA. And what that really means is that the medical professions will do their best to protect your personally identified information and your protected health information. And so where this comes into play with marketers, is, we’re so focused on collecting everything we can about customers and users and their behaviors and their demographics and, you know, their likes their dislikes, that we’re not really thinking about the bigger picture and what that means. Now, HIPAA, and all of these other regulations have been in place for at least 20 years in the healthcare space base. we’re now seeing this crossover into the marketing space. And so we actually saw an instance of this last week where someone we were working with a customer and there was enough identifying information, they, you know, took out names, they took out addresses, phone numbers, but there was still enough identifying information and health information about individuals in their database that it could have, had we not caught it been a HIPAA violation. And if you’ve ever experienced a HIPAA violation, it’s nasty is expensive. And it’s not something that you ever want to experience. And marketers are not equipped to go through these things, healthcare companies, it’s expected marketers, not so much.

Christopher Penn
But wait a minute, a man. So let’s say I’m a, a widget manufacturing company, I manufacture left hand, left hand to smoke shifters. HIPAA doesn’t apply to me, I’m not healthcare company.

Katie Robbert
And so that’s where there’s a lot of misunderstanding, HIPAA actually does apply to you. HIPAA applies to anyone who collects user data. So HIPAA is essentially protecting that health information, anything that it can indicate a medical condition, whether it’s physical or mental, any sort of a disability, anything about you, that you would go to the doctor for, or seek some sort of help for technically falls under HIPAA. Like what protected health information is this whole other episode, because there’s a lot that goes into it. But basically, think about it in terms of if I’m going to the doctor for it, it falls under HIPAA. And so in your instance, Chris, if you are a widget manufacturer, you’re probably trying to figure out the behaviors of your customers. So you want to know, age, gender, general location behaviors, likes dislikes. Now, if you are developing something that might they might use physically, you might want to know, do they have any physical limitations or those types of things that falls under that protected health information? And you can’t use that information? external, you can’t use that to market you can’t use that to publish research or anything? Because then that’s a HIPAA violation.

Christopher Penn
Well, you’re speaking as a as a former executive, I’m sure that can’t be that bad, right? Just pay a fine and move on with your day, right?

Katie Robbert
You’re hilarious. You know, it really depends. HIPAA violations can be pretty small, or pretty large, it depends on the amount of data. You know, this is not an instance where the government’s going to give you a slap on the wrist for not knowing any better. They’re not just going to turn a blind eye, but they go Well, you didn’t know you didn’t mean to do anything. Like it’s a pretty serious violation. And I’m not up to date on the amount of money that it would cost you in fines. But the fact that there are fines involved and paperwork involved, or just things you generally want to avoid, and to be honest, it can go a step further, where the individual whose information was violated, then also has a, they can take legal action against you. And that’s a whole other messy situation.

Christopher Penn
It’s interesting, because like with GDPR, one of the big reasons why everyone scrambled to implement that as quickly and as aggressively as they did was because there was a little clause saying senior executives can go to jail. Does that true also in HIPAA?

Katie Robbert
Um, you know, that’s a great question. I think that the there are criminal penalties, there are, I think the maximum is up to 10 years, with a fine of up to a quarter of a million dollars. And that’s where a fence is committed with the intent to sell transfer transfer, or us individually identifiable health information for commercial advantage, personal gain, or a malicious harm. And I want to be clear, they don’t differentiate between commercial advantage or malicious harm. So of course, not going to say, Oh, you were just trying to get a better marketing campaign, they’re going to say, hey, you did this thing, you’re not going to jail for 10 years, and you now owe us a quarter of a million dollars. So that’s the worst kind. So a HIPAA violation due to willful neglect, that is not corrected. So let’s say in your example, Chris, you know, you’re a widget manufacturer, and oops, you didn’t know. And then you continue to do the thing. You know, that’s $50,000 per violation with an annual maximum maximum of a million dollars until it’s corrected. So these are not small dollars that we’re talking about. This is something that could very easily bankrupt, a small, firm or mid sized company. Now, obviously, larger companies like they have more money to play with, but I don’t think that they’re in the business of wanting to just give away a million dollars for things that you can prevent.

Christopher Penn
Gotcha. So let’s go back to the this marketing and this customer data thing. Let’s say that we know as marketers, okay, so we shouldn’t collect obvious health care data, right? Like, don’t send me your blood pressure, right? We’re not going to make your heart rate part of the the need to fill out this ebook form to attend a webinar or whatever. What it sounds like there are still risks, even with things that are not explicitly healthcare data, because there’s other categories of protected information. So how do we how do we think about this? How do we mitigate this so that we’re not setting ourselves up for either HIPAA or GDPR? Or CPA violations?

Katie Robbert
Well, you know, it really it starts with What’s your goal? What is it that you’re trying to accomplish? Everything you do, should always start with that plan? And what’s the question you’re trying to answer? So if you do find yourself needing to ask consumers about, you know, have you ever saw medical treatment? Have you ever seen a therapist, whatever the question is, you need to collect that data in such a way where it is encrypted, and it cannot be reverse engineered, you know, you can’t like the database needs to be fully protected, you can’t hack into it. And the customer data, the individual record, the demographic information needs to be kept separate from the identifying information, all of those different characteristics, those markers that we just talked about. And then when you do put it together, it’s the anything that can identify the person also, that needs to be encrypted and encoded. And you can’t be you can’t have the ability to reverse engineer the demographic information. So for example, let’s say that, you know, you collected data about me, you know, a, between the ages of 35 and 45, Caucasian female lives in Massachusetts, if I lived in a town with less than a certain amount of people in a very small population, you wouldn’t be able to put any of my identifying information in there, because it, they could very easily figure out who I am based on other characteristics. So it is actually I wish it were simple, but it is pretty complicated. The rules around protected health information and personally identifiable information. There are many there are complicated, but they are very clear cut. And the ways in which you can prevent against any of these violations are fairly straightforward, like so if you’re collecting information on your customers, just don’t do anything dumb with it, don’t publish names, addresses, phone numbers, don’t publish health information, and you can’t use it. to market to these people. You can’t use that health information.

Christopher Penn
What about proxy variables, though, so in machine learning, one of the things we can do is infer quite a bit of information from other information if we have labeled training data. So for example, let’s say that everybody who likes a niacin and the movie, the blue is the warmest color and happiness like tacos. For whatever reason those three identifiers predict the likelihood that somebody is LGBT q with about 99% accuracy, and further predicts that they are likely to have an allergy to shellfish. Do those proxy variables then become protected health information if the machine learning model has inferred that with a high degree of, of statistical strength?

Katie Robbert
Yes, it does. And now, here’s the thing, I apologize if I have been clear, you can collect this information, it’s what you do with it, that makes a difference. So if you’re using machine learning model, to infer all of these characteristics about your audience base, that’s fine. It has to stay internal, it has to stay under lock and key and encrypted. You can’t publish it externally, and you can’t use it to mark you can’t say, okay, we are going to create a campaign around these LGBT Q, you know, ages 35 to 45, who have a shellfish, shellfish allergy, who also likes the three things that you named, like that’s using that protected health information for commercial advantage.

Christopher Penn
So when people talk about one view of the customer, and and you know, customer data platforms and things, then are they thinking about this? are they thinking in terms of I might be extremely Apple waiting, sensitive health information or sensitive financial information?

Katie Robbert
My an educated guess is No, they are not thinking about it. Because we see, you know, working in the healthcare space, there was this constant microscope, and you were always thinking about it, when I stepped out of that space. People just didn’t even know how to protect their information, or how to protect their privacy or what to do with the data they collect in its living in a spreadsheet that shared on a cloud server somewhere. And, you know, it’s, I would say, No, people cdp’s are not necessarily thinking about it, or the people who design it, have thought about it. But then when you give it over to someone else to use, it’s not something that people are trained on very well.

Christopher Penn
How do you do marketing then in, in like healthcare, if you’re not allowed to use health data to market like, if you have a shellfish allergy? And I have, I don’t know, shellfish allergy mitigation pills? How do I market to you if I can’t use that data?

Katie Robbert
It’s tricky. If you’ve ever tried to market in the pharmaceutical or the healthcare space, you know that almost every digital advertising platform will disapprove your ads or throw back arrows and say, You can’t do this. And essentially, it comes down to targeting those individuals, you can do sort of the broad strokes. And even then, the language that you can use in healthcare advertising is pretty limited. It’s one of the biggest issues within pharmaceutical sales. So you’ll see those, you know, three or four page in a print magazine, but it’s not targeting any one individual person, you can’t focus on, you know, an individual, the way that you can with a digital marketing campaign. So you rarely ever see pharmaceutical advertising on like social media platform forums, for example, because there’s just so many regulations and rules against it the same with health care. So if you’re, you know, you need to sort of get up to date on what those words are, that you can and can’t use, how you can target those folks. And it is it’s really limited, it’s really tricky. And it’s all in the name of protecting these individuals.

Christopher Penn
Gotcha. So if I was a company trying to market, you know, shellfish allergy pills, and I found a Facebook group, you know, called, I hate shrimp. Could I or could I not advertise to that groups and say, self identified as people who don’t like shrimp?

Katie Robbert
You know, I honestly don’t know. I’m sure that there would be rules within the group saying, you know, no advertisements, no solicitation, and you might immediately get bumped from the group. If you do try to push your shellfish allergy medication on them, because their people are so protective of that person health information.

Christopher Penn
Gotcha. Okay, so let’s say that I’m a marketer, I’ve got this great database. And now after hearing this, I am thinking, Hmm, I do have things that could be construed as health information to what am i next steps? What should I be doing?

Katie Robbert
Well, if you have a database that potentially has protected health information, or personally identifiable, if viable information, you need to make sure it’s encrypted. So if it’s just living in a spreadsheet, on someone’s Google Drive, or somewhere, that’s a terrible idea, that is immediately hackable, and that is technically a HIPAA violation. So make sure it’s secured somewhere, like a sequel database, encrypted, make sure it has a password. Those are sort of the basics, like use your brain, put a password on it.

Christopher Penn
Okay, but but then, what else should I be thinking of in terms of how I use that data yet? Let’s say, you know, I did collect by accident, you know, so seem to have attracted in the audience, people who have shellfish allergies?

Katie Robbert
Well, so your first question is, do you need that particular data point? Do you need to know that your customers have shellfish allergies, or is that just something that you happen to collect, if it’s something you don’t need, you should immediately delete it and erase it from your database, because it’s not going to help you if it is something that you need. And you also happen to have information that could identify any one individual person, that data needs to be kept separate. So the demographics and the behavioral data need to be kept in two separate places, they can all be in a sequel database, but different tables, different passwords, it just needs to be kept apart, because really comes down to the core issue of you should not be able to identify any one individual person and all of their physical and mental ailments.

Christopher Penn
Okay? What about stuff like email mark, like I publish a newsletter every week? And inevitably, there are all these different types of links from different kinds of topics and things like that. And every now and again, there are health topics in there. Does the click through data for health specific articles constitute protected health information? Because I have personally identifying information I know who the person is, and I know what they clicked on?

Katie Robbert
Not necessarily because they have, there is no information that says that they have that individual ailment.

Christopher Penn
Okay, gotcha. But it does say that they always click on mental health articles, for example,

Katie Robbert
right. And so just it comes back to what is your goal of using that information. So if it’s just, they always click on those articles, you know, that’s one thing, they may have a family member, or they may just be curious, or they may be studying to get their PhD, or whatever the case is, it’s when they actually check a box to say, I have this ailment, this is part of my medical treatment, that it becomes it turns over into protected health information. I am fascinated by psychology personally. So I always read mental health articles, because the way the human brain works is really, really interesting. That’s not protected health information. That’s just something that I happen to be interested in. But if my physical and mental health records, somehow make it into your email marketing list, that is a HIPAA violation.

Christopher Penn
What about Then, can a system inadvertently become contaminated? So let’s say you have we have a contact form on our website, and somebody you know, fills out the form says, By the way, you know, I have a shellfish allergy. And I’m looking for a data science specialist team that can you know, can avoid even mentioning shellfish in their data? Because I might reaction to it. But they have now volunteered personal health information into our contact form on our website, does that now inadvertently convert that data into p dry?

Katie Robbert
Yes, yes, it does. And so and these are challenges that, you know, I don’t think the marketing industry is really well prepared for and so what you would need to do in that situation is, you know, contact the person, let them know, you know, thank you for your information. What you’ve shared with us is protected health information, you know, so we are going to go ahead and remove that information from our records in our database. And just if it goes beyond that, then you know, there’s a lot of other steps that you need to take, including potentially reporting a HIPAA violation, if you are proactively reporting, the fines, and the violations are, they go a lot easier on you. But if you’re found out, that’s when you start to get into the million dollar 10 year jail sentence, yeah, you need to immediately nip it in the bud. You can’t say I hope nobody sees this.

Christopher Penn
It sounds like then that everyone who collecting open ended data from customers or potential customers has to have has a built into their processes, like almost like a compliance check to make sure that, you know, especially if you’re going to take like you’re like one of things we often talk about is take your customer service inbox, take your CRM data, and use it took your mind that data for insights for you know, building new marketing, things like that, it sounds like a part of that process has to be first validate that you don’t have disallowed information in there.

Katie Robbert
That’s exactly it. And you know, what you’re describing is what we’re always encouraging people to have is that data governance plan. And so, you know, we as marketers usually think that the data that we’re collecting is light and fluffy, its social media likes, it’s those types of things, but it can get very heavy very quickly, it can get very valuable. And so, you know, you need to be building in that process of is that the data that I’m collecting protected health information? Is it personally identifiable, personally identifiable information, I clearly need more coffee is that sensitive information, you know, and if you don’t know what those things are, Okay, first and foremost, Wikipedia is a great resource for all of that stuff it breaks breaks it down very easily. But we’ll put links into the show notes for each one of these things, including what HIPAA stands for, and what it means to violate HIPAA.

Christopher Penn
Hang on, wait a second. So suppose I’m collecting tweets about SXSW, your huge event mega huge event. But there’s a panel there about, say, shellfish allergy. And a bunch of people may who may or may not be at the show, use the hashtag, use that use that panels hashtag and talk about their personal allergies. Does that mean now my social media monitoring system also has protected health information it because I’ve collected those tweets and say, I have a shellfish allergy GM, I still eat shrimp every day.

Katie Robbert
It depends on what you’re doing with that data. If you are exploiting that data and pulling it into your own personal systems, then you need to do your own encryption and protection. If that information is out there, that social media monitoring tool is responsible for that data, you know, you seeing something on the internet, on Twitter on whatever is not a HIPAA violation.

Christopher Penn
But if I collect that data, if I download the data from my social media monitoring tool, I now have to protect that data.

Unknown Speaker
That’s correct, thank God, I encrypt everything.

Katie Robbert
But I would guarantee a lot of marketing teams, especially smaller shops, don’t even know where to start with even encrypting their data when what that means. So I, you know, there are a lot of steps to it. And so it’s the conversation is not to scare marketers away from collecting behavioral data about their customers, is to start to challenge I want you to be a little bit more responsible with how you’re collecting the data, how you’re storing the data, and how you’re using the data. And so really think about it like this. You cannot use the data that you’re collecting to target any one individual person, even though that’s probably what you want to be able to do. So if I know, Chris Penn lives in this area is of this ethnicity in this age, and he happens to be a psychopath, because it’s identified in his mental health profile, can’t then create a campaign targeting this one individual Chris Penn, about an anti psychotic medication? As much as he may need it, it may help him. I cannot do that.

Christopher Penn
All right. So to wrap up, it sounds like at every point in data that deals with individual people, regardless, probably is the best practice regardless whether contains protected health information or not. You should be security, HB encrypting it, you should be preventing it from misuse. And you should be declaring the intense and the what you are going to do with the data before you start collecting it so that you’re very clear. This is what we will and we will not do. We’ll have links in the show notes to summarize some of this stuff. And I think, Katie, we’re going to see in here a lot more from you on this in the next year or so. Right?

Katie Robbert
Yeah, there’s a lot to unpack here. This was a very high level broad stroke introduction. But there’s a lot that goes into this. So the goal is to really educate marketers on how to be really responsible with their customer data.

Christopher Penn
So stay tuned to the trust insights blog and to this podcast. If you haven’t subscribed, please subscribe. Just go to trust insights.ai slash podcast for more instructions there and we will talk to you next time. Take care


Need help with your marketing AI and analytics?

You might also enjoy:

Get unique data, analysis, and perspectives on analytics, insights, machine learning, marketing, and AI in the weekly Trust Insights newsletter, INBOX INSIGHTS. Subscribe now for free; new issues every Wednesday!

Click here to subscribe now »

Want to learn more about data, analytics, and insights? Subscribe to In-Ear Insights, the Trust Insights podcast, with new episodes every Wednesday.

Leave a Reply

Your email address will not be published. Required fields are marked *

Pin It on Pinterest

Share This