Cory Munchbach has spent her career on the cutting edge of marketing technology and brings years working with Fortune 500 clients from various industries to BlueConic. Prior to BluConic, she was an analyst at Forrester Research where she covered business and consumer technology trends and the fast-moving marketing tech landscape. A sought-after speaker and industry voice, Cory’s work has been featured in Financial Times, Forbes, Raconteur, AdExchanger, The Drum, Venture Beat, Wired, AdAge, and Adweek. A life-long Bostonian, Cory has a bachelor’s degree in political science from Boston College and spends a considerable amount of her non-work hours on various volunteer and philanthropic initiatives in the greater Boston community.
Richie helps individuals and organizations get better at using data and AI. He's been a data scientist since before it was called data science, and has written two books and created many DataCamp courses on the subject. He is a host of the DataFramed podcast, and runs DataCamp's webinar program.
I can't stress enough that I think that privacy as a competitive differentiator is really something to watch. I'm excited about the brands that I see who are baking that in from scratch and really making that a part of what they do. I think that's going to be pretty remarkable to watch and to see how that unfolds
Technical marketers, technical data engineers are really starting to show up in marketing or more business oriented parts of the organization in ways we haven't seen before. I was just having a conversation with a partner of ours recently who talked about five years ago, they didn't have any data skills in their marketing organization, and now they have data engineers. They have sort of data quality folks. They have folks with expertise whether it's knowing SQL or other kind of coding languages on the really technical side, certainly more data scientists, folks who at least have a passing familiarity with Python and things like that. That's gonna be very much more a valuable addition to a marketing organization.
And then on the customer analytics side, there's a lot of tools that require you to, again, passing knowledge at least of HTML, for example, if you're working on the website of things. So, you know, there's a core skill set that the more you have of it, whether it's on the data science, analytics, or in the kind of harder core types of disciplines like SQL and things like that, there's a nearly unquenchable demand for that on the business side. As marketing tries to figure out how to bring these things closer and closer together, those who have typically lived in IT or in agencies before will bring those kinds of things in-house. Being familiar with all the right tools that you need, the analytics platforms, certainly CDP, web tools, like those are the core competencies of a good martech stack and having knowledge of how those work and being able to sort of have at least like entry-level marketing ops kind of chops will set apart candidates now.
There is a growing demand for data engineers, data quality specialists, and professionals with expertise in SQL, Python, and other technical skills within marketing teams. Enhancing these skills can be crucial for staying relevant and effective in the evolving landscape of marketing technology.
Data clean rooms enable organizations to share insights without compromising data privacy or ownership. Understanding and utilizing these tools can be essential for navigating modern data privacy challenges.
When conducting experiments in marketing and customer analytics, align them with specific business outcomes. This approach ensures that your efforts contribute meaningfully to the organization's goals and are not just theoretical exercises.
Richie Cotton: Welcome to DataFramed. This is Richie. Using customer and marketing data well can dramatically improve the customer experience, as well as making your company more money, magnetizing people to spend more, and even unlocking new markets. It's a fascinating time to be in the customer analytics space. On the one hand, improved tooling and increased data savviness of the marketing teams mean that much more sophisticated analyses are becoming mainstream.
On the other hand, a decade of scandals around misuse of customer data has led to stricter laws and increased scrutiny. That means that to be successful, you need to be much more mindful about what you do with your customer data and how you do it. Helping us thread the needle to customer data success is Cory Munchbach, the CEO of the customer data platform BlueConic.
Cory has spent the last decade helping companies achieve marketing analytics success while respecting their customers data privacy. She's at the forefront of how this space is changing and is a font of practical wisdom, so I'm looking forward to hearing her advice. Hi, Cory. Thank you for joining us on the show.
Cory Munchbach: So glad to be here. Thanks for having
Richie Cotton: Brilliant. So, to begin with can you just give me some examples of when data has been used to have some kind of success story where it's had an impact on the customer experience?
Cory Munchbach: Yeah, I mean, look, we could do this the whole podcast of all of... See more
But one of the really cool things that they've been working on is this idea of Mattel creations. Which is a digital channel and it's really focused on sort of the collector and enthusiast consumer base for them So sometimes adults not necessarily kids who have been big fans for a long time And they used a lot of the first party data that they had to understand who these collectors are, how they engage with the Mattel brands, and basically built out this whole new channel, Mattel Creations, using, using that insight.
So that's an incredibly cool one just sort of the innovation and diversification that they were able to bring on because of data. On a totally other type of industry and another part of the world for that matter Talia is a telco in the Nordics, covers all of the Nordic countries they are doing a whole bunch of work around this idea of personalization at scale.
So they have a much more prescribed life cycle, right? The telco typically has renewal moments and some of those more predictive aspects of their customer journey. And they use data about their consumers to understand where they are in that life cycle. They make sure their segmentation is dynamic based on sort of the signals that they're getting as the consumer interacts with Talia and uses their different services on digital channels.
Consent being a big piece of that given that they're in Europe in particular I mean using all of that insight and scoring to make sure that they Whether it's marketing customer service are putting the right offers in front of the customer depending on where their journey is. So Two very different examples, but two of my favorites just in terms of how powerful having that first party data has been for both of those organizations.
Richie Cotton: Yeah. So really, if you don't understand like what your customers doing or what they want, then you can't help them. So Now, you mentioned the idea of first party data, just so everyone understands what's going on. Can you talk about what that means and what the different types of data are?
Cory Munchbach: Yeah, absolutely. So, I think this is an interesting time to have this conversation. So, there is this concept, I think, even of, like, zero party data, which is essentially this idea that it's data, a consumer directly and explicitly provides to an organization. First party data is one click above that.
It may be implicit or explicit data that the consumer provides to a brand directly. So it's really that one to organization, that consumer to organization or customer to organization type of data. Second party is that same first party data but shared between two organizations. So potentially if you are a consumer packaged goods, for example, company, you're sharing directly data with a retailer.
But again, just between those two organizations only, that would be second party. And then third party data, the one that is kind of on its way out and but has been one that has really driven a lot of marketing and advertising for quite a while. That's data coming from lots of different parts of the world whether it's the internet, other sources brought together by data brokers or ad tech companies.
aggregated into generally some kind of anonymized type of database that you can then go in and buy data directly or buy ads or other things against that data. So that's kind of the spectrum of, again, consumer explicit, given consent, direct actionability, all the way up to kind of further remove in that third party ecosystem.
Richie Cotton: Okay. So quite the spread. And then you mentioned before as well that all this data gets used for things like marketing and for customer service. Can you talk me through some examples of how customer data is used at each stage of the customer life cycle?
Cory Munchbach: Yeah, absolutely. I mean, look, this is kind of the power of the data is that it can be used across that whole life cycle. So when you think about acquisitions or top of funnel or early life cycle, however you want to think about that that can be when you don't even know who somebody is yet more of that anonymous kind of One to many type of framing but being able to be more intuitive about the type of offers that you might put in front of someone or If you're in media getting them to sign up for a newsletter or things like that Maybe agree to sign up for news Different types of offers if you're in retail or manufacturing but that precision up front with the targeting to kind of move someone into the life cycle is really common.
And then you can build on that in the more the engagement part, sort of mid funnel or mid life cycle again, however you want to think about that. But that can be The fact that you've come back many times we know what you've purchased before we know what you've read before for example So how do we make sure that we're giving you?
New things to engage with or better suggestions based on your previous behaviors And then true through to conversion So making sure that again you've already bought this thing or you already have one of them If I use the talia example, right? It's different if you're a net new customer versus you're renewing Those are different conversion types of moments How do we make sure that we're streamlining you to actually do that in the right way?
Based on the fact that we know who you are and where you are in that journey. And then retention, I think is probably the one sort of the retention loyalty side of things is where. Data has the can have the most power in a lot of ways because it potentially can serve as kind of a bridge between those conversion moments and when you hopefully have another conversion moment, whatever that looks like for your business, but being able to engage more intelligently, smarter email, smarter advertising, smarter follow ups, all of those kind of things that allow you to kind of Sustain a conversation with the consumer in the right way really can have an impact on that sort of post purchase retention loyalty side.
And as I mentioned, the example Italia to starting to move beyond marketing and think about how can this be used in support or customer success and those kind of functions. Outside of the core marketing with the reality that every interaction with a brand is part of that experience. So just because marketing doesn't necessarily own it being able to make sure that that is smarter and better and more aligned with the brand is in all totality more effective right so that we can think about it that way So really the data what's most critical is that at each phase?
You're building on the data from other parts of the phases, right? The other parts of the life cycles. One of the things that's been so historically problematic for a lot of organizations is that each team or teams that owns different parts of that journey are operating on different data. They're not talking to each other.
It's not seamless. So being able to create that continuity of the data just to support how a customer moves through that life cycle is so critical and then tailoring it by the life cycle objectives is kind of the second part of that.
Richie Cotton: Certainly avoiding data silos to simply well, it's a problem that every company faces. So I definitely want to talk about that more in a moment. Before we get to that can you give us some examples like what sorts of customer data tend to be collected? Is it just name and email or what are we talking about here?
Cory Munchbach: For sure. I mean, first party data can be anything, really. And I think this also gets us to one of the critical changes in the marketing world right now is that it needs to be consented. It needs to be thoughtfully collected. So before we even get into types, whatever it is, this emphasis starting a lot with GDPR, but also now moving into more of the California laws and other places that are enforcing just a better framework around Why someone may be asking for the data and what the purpose of it is for the brand.
So cultivating that trust is so critical for marketers now preceding whatever it is that you're going to collect. And then when it comes to the type of data, it can be exactly what you're describing. More explicit stuff about who I am, kind of my demographic information, but it can also be transactions.
It can be behavioral data from the web or mobile, other places like that. Sort of those signals that you can get from engagement. The devices that you're using if you have consented and for what purposes, if you've gone to events as the case may be, right? You've shown up in person, all of that can be customer data that can be collected and unified in this way, but it's just so critical to stress that.
More and more. It's important for marketers to be thinking about just because you can doesn't mean that you should and really thinking through. Why do I need this? What might I use it for? And ultimately, how does it respect and serve the consumer and the business goals that I'm trying to get to? I think we're still in our early phases of the way that that's been adopted, but it's absolutely where we're going And we're going to see more and more of that from brands going forward.
Richie Cotton: So that's really interesting because you mentioned that you got sort of GDPR, you've got CCPA, all these sort of laws around privacy and, companies have to think a bit more carefully about what they're collecting and it does feel like maybe what, like five or six years ago, there's a point where you wanted more customers, Google and Facebook knew everything about you.
It's like, okay, we don't think about you. Which, which people do you want? And that seems to have changed. So can you talk a bit more about, like, how these sort of privacy laws are changing what kind of data does get collected?
Cory Munchbach: Yeah, look, I think we're still very early days of this and in some respects we're in a challenging kind of teenage tweener phase, I think of some of this. So on the one hand, GDPR was such an important forcing function back in 2018 to put this general data protection regulation around what it is that can be done.
What I would argue is the problem is that that is just now resulted in like basically the internet being harder to use with all the pop up boxes these sort of blanket accepts like you know, that was Not the intention of those regulations by any means just to make the internet even more annoying to be on.
But what it has done is sparked a really important conversation and these kind of look alike laws, right, that you mentioned, like the CCPA and others. So it's moving us forward. It's just messy and sloppy right now. I think a lot of the things that have led to this, one is we saw a ton of more just kind of high profile data breaches over the last many years.
Whether it was meta having those, but also just consumer brands and just people being generally more aware about the security of their kind of personal information. You add in the regulatory changes that we talked about. Then you combine the third party cookie deprecation, so web browsers, everybody, whether it started with Safari, Firefox, Chrome, following suit, really has just changed the expectations for how we engage in the internet.
And all of that was true and necessary before we started to get into even just the last 12 to 18 months. where you see a much, much more stringent, effective algorithms on top of that data. So there were already concerns about how the data was being collected. Now it's also even more scrutiny on not just the collection and the source of it, but how is it being used and how is it feeding algorithms?
How is it making recommendations? What impact is that having? And those kind of both components, the unification collection of the data. And the kind of use and activation of it. All of that has put a pretty damning light on Facebook or Meta, Google and some of those bigger companies that have done that basically without any scrutiny at all.
And when people started to look, it was not good behind the scenes. And so all of that is, is kind of demanding this broader reckoning that we're seeing.
Richie Cotton: I think there's definitely more awareness now from consumers about what's happening to their data. So how can companies deal with this new environment where consumers do care a bit? And there are these regulations. So what do companies need to do to make sure they're complying and make sure they're doing right by their customers?
Cory Munchbach: So in sort of a hierarchy of compliance, there's just This is what the laws say, where are you operating, where are you engaging with consumers? Can you meet the letter of the law wherever you are operating? there's so many tools now out there to do that. If this is still an open question for you, you've got a bigger problem on your
Richie Cotton: yeah, if you're doing a legal customer data collection, probably not
Cory Munchbach: pretty much like I'm not a lawyer, but I just think that maybe you should probably hang up the podcast right now and call your general counsel because that that would be a concern. But the
thing is that like while everyone was rushing to comply with the letter of the law, this is where you then end
up with the chaos that we now have all these
stupid pop up boxes, just very thoughtless, not customer friendly. Types of
disclaimers, right? It was, it was literally
just that like good enough is good enough. What I think is so, exciting or at least a huge opportunity to come out of this is in the same way that we had this phase with digital in the early two thousands where everyone just had a shitty website or an app and it was just like, I checked the box.
I did it go, tell the CEO I did it. Privacy can be similar. A really thoughtful, intentional, privacy first orientation can be, and I think for a lot of brands, will be a huge competitive differentiator over this next kind of 10, 20 year period. And what I mean by that is thinking about how does privacy align with the value of your brand?
What is the role in this? How do you not just kind of weaponize some of this as we've seen some brands do, but otherwise, like, How does this fit into the overall experience? And I think when you look at brands that have done this well, it's this idea of kind of a value exchange. It's not just giving up the data for the sake of it.
People can understand what they're doing and then safeguarding those data when you collect it, that's going to be also really critical, right? And, we just saw 23andMe had a data hack so this stuff, whether you are just the consumer facing side of it, or recognizing like how you protect data, this is an asset this is a currency of the modern economy.
And so it needs to be treated as such in a thoughtful, intentional way. Again, we're still early days, I think, and really having a super strong playbook for this. But I would argue that within the next couple of years, maximum, we're really going to see the brands that sees this as the opportunity that I think it is.
putting distance between themselves and those who fail to do so, and certainly between those who screw it up or just disregard it altogether.
Richie Cotton: Yeah. So you said something really interesting that ideally you want customers to get something in exchange for giving you some data. So have you got any examples of when that happens? Like, I suppose 23andMe example is like, it's a fairly obvious trade. You give, you give data about your genetics and you get some information about your family back, but do you have any more examples like this?
Cory Munchbach: Yeah. Well, I think that really the most clear example of this is, is more in the wearables space. So, whether you're a Whoop or an Apple Watch or any of these others where you are able to track your steps, your physical activity, your heart rate any of these things that sort of data is incredibly personal, but people find it beneficial for them, whatever it means for health reasons, for their own entertainment, weight loss, whatever it might be.
There's clearly a willingness to do that well. I would even go maybe more simple than that is media and publishing companies who, Really effectively harness what it is that you read on their website and refine the web experience or the mobile experience based on your content consumption. That's not super land like groundbreaking, but it is really, really impactful when done.
Well, you're cutting through the noise. You're getting insights and information in front of people in a way that they care about, and it is super useful to them. And so that's a very small example, but I think one that we often underestimate. How impactful that can be and getting people to come back, for example, right, to be able to create that ongoing engagement.
If you contrast that with media companies who basically took every square inch of a website and covered it with advertising that may have gotten them short bursts and hit of revenue, but has dramatically declined because the consumer hates it and they're going to go find alternatives. And so I think that is what I mean by the value exchange is just understanding what is this experience supposed to be like for the consumer?
How do we make sure that that is designed in such a way that sets up the longevity of a relationship, not just maximizing the most dollars per interaction and instead thinking about the broader lifetime value and so I think again wearables are really like hardcore kind of sexy one. That's easy to talk about but you look at media you look at netflix with their recommendations algorithm, right things like that that are not necessarily So kind of, sci fi, but are hugely impactful from a revenue and long term relationship standpoint, and those are just a couple that kind of come to mind for me.
Richie Cotton: Yeah, these are all great examples. And do you have any more advice around like, where the sweet spot is? Like, what are the sort of low hanging fruit in general for improving customer experience based on these sort of data experiences?
Cory Munchbach: Yeah, so I've been doing this stuff for, quite a while, and the common theme that I would say among the companies that are Very successful in this type of harnessing data and using it in an effective way are the ones who think about it much more as a discipline of their organization. They think about it not as getting some big personalization program launched or kind of in a specific milestone.
But rather in how do we transform our organization to just be better at using consumer data in a way that is effective and safe and productive for our business. And so that is the least sexy answer, but it is the one that works over a long period of time. And, and right, you can kind of analogize
that with like the person who just shows up and works out every single day and is super healthy and kind of builds that over time for themselves.
First, the person who like shows up one time.
gets their one rep max in and like never shows up again, right? It's like, it's a totally different mindset. and there may be a few people who can get away with
that works for them. But by and large, it's incremental. It's a mentality of incrementalism.
It's a mentality of how do we get 1 percent better at this every single day that pays off. in spades over a longer period of time. The problem being that everyone wants the more instant gratification. Executives want to see what, what happened with this investment that I made. And I think that short term ism versus the longer term ism creates understandable challenges.
get that. But if you do sort of step back and look at it from that slightly longer term point of view that the trick is in the incremental, the building on your success, the testing and learning, really building, building, building, building, as opposed to trying to kind of just do everything in one fell swoop.
Richie Cotton: I do love that idea. That's a, well, I suppose it's like an agile approach just going little bit at a time. And I know like personalization has just been talked about a lot in this last year because Generative AI makes it sort of cheap to create content, and then it does provide more personalized experiences.
So can you maybe talk a bit about how generative AI is changing things in terms of customer experience?
Cory Munchbach: So there's a couple pieces to this for me. One is do think there is an enormous step that can be taken because we've long had the conversation with our customers, for example, about how well, can I make a segment of one? I want to be able to do one to one marketing and i'm like you can do whatever you want Do you have copy and content and visuals?
To support that kind of one to one marketing and of course the answer is always no. However, generative AI may be something that allows us to take that massive leap forward as far as creating those one to one or certainly one to fewer. Engagements and personalization that we haven't seen before, so I do think that will be a big piece of this.
Another aspect of it and again, similar to what I just said about the incremental 1 percent better type of transformation, I feel the same about generative AI in the sense that it's going to help automate. The routine, the rote, the stuff that just takes a lot of time, companies that harness it for the less sexy, repetitive activities and free up your brilliant marketers and digital teams to be creative and to think new big thoughts and to try new things that only they can try.
Those are going to be the folks that win versus the ones who maybe come out really hot out of the gate with some. Cool generative AI kind of concept, but it's not being built into how their organization operates. And I think that's true again for the people as well as for the content side of the house.
That figuring out the right way to bake this in with intention and being deliberate is going to pay off.
Much more likely in a couple of years, in a big, big, way versus ones who kind of claim to have it figured out right now. But probably I would
expect flame out because it's not embedded.
It's not structural. and this type of technology has the potential to make those kind of big changes, but it's got to be treated
in a structural way in my view.
Richie Cotton: Absolutely. That does sound important. And I suppose related to this idea is we ought to talk about customer segmentation, because that seems like, that's the in between step from we're doing everything just for everything the same for all of our customers to we're having everything personalized.
And so do you maybe have any thoughts on where that fits in?
Cory Munchbach: I mean, looks, I think you're exactly right. Segmentation is to me kind of where all of the magic happens. It's where all the ingredients that the data is come together and create something unique and special. it's long been, I think, the part that is most fascinating to me about the work that we do is how our customers think about their own consumer segmentation and what's important to them and how that varies by industry and by organization on all of that.
I think a thoughtful segmentation strategy is really where the rubber meets the road, right? It's the intersection of the data and the activation, exactly as you say. And so that can both also have an AI impact in automating. Some of that testing and being able to hold out test groups, things of that nature.
But also figuring out how can we start to do more complex testing of different segment combinations with different content combinations, right? Much more true kind of multivariate testing at a scale that we maybe haven't seen before. But really it does. It all comes down to who are these cohorts that we're interacting with?
At what altitude are we considering our consumers? I think one of the discussions we often have is the difference between kind of a segmentation analysis, which might be very familiar to folks working in media and advertising, but those are essentially personas versus actual dynamic segments that people are moving in and out of.
We interact with these people in a particular way. And then same thing true more for cohorts, which is more of an analytical way of thinking about the groupings. So in marketing and advertising and data, I feel like we have all these different kind of ways of thinking about groups of consumers.
So getting that, ontology right and what we're actually discussing in our world, is more about the activation and kind of creating a dynamic type of segment. Everything else still has a huge purpose. They should all be kind of philosophically connected. But in our world, it is about how do we take this group of people and make sure we're interacting with them in the right intelligent way.
without that, you end up with the noise of all the data and the aspiration of all the activation. But without the segmentation in the middle, you can't bring those things together in the right ways. I
Richie Cotton: That's fascinating. And it sounds like what you're saying that Different parts of your company are going to group your users in different ways. So maybe like the product team is going to slice and dice your customers in a different way to your marketing team or whatever. Is that a good thing or a bad thing?
I'm not quite sure.
Cory Munchbach: think it really depends, honestly. So the way I would use that as an example to kind of pull that thread further, if we have a pool of a million people, right, who have interacted with our brand in 2023, just for the sake of argument. I don't know that there's anything wrong with necessarily looking at that a million from the lens of who bought product A versus product B.
I also think there's nothing wrong with looking at it from the lens of who engaged with us in social versus who didn't, for the sake of argument, right? Really simplistic things. What's really interesting is, though, what difference does it make? And being able to look at that and say, is there anything intrinsically more valuable or more specific or unique about those different groups?
And then furthermore, what if we combine them? What about looking at people who brought product A and engage with us in social? What can that tell us? So how does it really progress the way we can think about engaging? In some cases, I think what's super invaluable in being able to have those conversations and having the data all together, which is where, the magic, I think, of our platform has come into play, is we can look at the overlap.
Is there any difference? One of my favorite examples was university, actually, and they were using the same, they'd always been using different data sets, they finally consolidated it, and the development office and admissions were looking at the same kind of student data, but Proxy here for consumers and what they found is that basically they had almost overlapping segmentation, even though they were, they were calling those groups of people totally different things and being able to look at them live and say, Oh, this is actually 80 percent the same audience.
How do we now want to talk to them in a different way instead of, bombarding them from two different places, but from one brand one university. How do we get smarter about that? And so those are examples of finding those efficiencies. You asked a question about, you know, cool examples. These are the unsexy ones, but have massive operational implications, right?
You, if you're spending money to spend, send those emails or, or pay for those audiences in paid media. That's a lot of money you're throwing out because you're you're essentially doing it in a redundant way or kind of harassing the consumer. And so even though you may have the different definitions, what I think is critical is to have maybe the same data asset that you're
working off of, but that you can
then go assess those individuals through all of your different lenses and then make the right decision
for the business.
But that goes to Yes, you need to break down the
silo, but you also need to have the kind of operational guidelines and parameters to work with the data
in that way. And that can be, that certainly can generate a
lot of friction and frustration, if not everyone maybe is in agreement about how we should look at it.
The fact that it serves that forcing function is exactly why it's ultimately very valuable.
Richie Cotton: Absolutely. I have to say Email campaigns is something I normally try and stay away from because it does seem like it gets incredibly complicated on how you divide people up and how you have like a different cadence for like which emails you send to which people at different times. And then how do you avoid spamming people?
So actually, I don't know whether you have any advice on that. While you're here, can you explain it?
Cory Munchbach: know, this is again, I do think it comes back to having one source of truth on the data side of things, and then being able to do be able to see, does it make a difference that you cut the segment in half and actually you're sending two different emails or is the performance essentially the same if you keep everybody together?
But I think that again, that's where the sort of the magic is in inches with a lot of this. And for most brands that are reasonably progressive on, on their digital marketing and all these sort of areas. There, the, the magic is going to be in the inches for them. Very few have greenfield ahead of them on some of these tactics.
We're all doing a lot of the same commoditized things, but what's not commoditized is your audience. What's not commoditized is your consumers. So really that's the asset that you have to work with and it's up to you as a brand. To make sure you work through that challenge and those difficulties in order to capitalize on that to make it a differentiator.
Otherwise, exactly as you say. You're just going to keep kind of throwing things at a wall and, and hoping that you get a different outcome. but in all likelihood, you're really not. the data is the magic here.
Richie Cotton: Okay, so, it sounds like we've got two things to talk about. There's data quality and the experimentation. Let's go with experimentation first. It's probably the easier thing to discuss. So talk about like, I know it's an odd situation where experimentation is the easy bit. So, yeah, do you want to talk about, like, what kind of experiments you might want to do in order to see that you're getting an improvement with your customers?
Cory Munchbach: Sure. So, I mean, level one for this is simply to assess, okay, these are the segments that we are working with today. Whatever that baseline is. And make sure you actually have a baseline of performance with whatever it is you're doing. Whatever marketing programs you're running. However you are defining success based on the audience that you have, make sure you have an actual baseline.
Then, you can start adding in layers from there. I say that only because I am it never ceases to amaze me how many do not actually have a baseline to work off of to know if the experiments are having an impact. So, start there. Second, assuming you have that then, is just to look at what do you think The highest level variables are to your business, right?
What are the business outcomes you are trying to drive? You may be able to find some crazy niche hyper performing segment, but if those people don't spend money with you or aren't going to ultimately help your business move forward, Then just because they click on your emails or whatever, it doesn't necessarily help.
So it's, it's really, this is where marketers, I think often don't necessarily get enough credit, but these are programs designed to help grow the business, whether it's grow subscriptions, grow transactions, but being really clear about the business outcome that we're trying to drive. And then your experiments should be compared to that.
You cannot just compare experiments against each other. You need to experiment against what is the business goal that we're trying to accomplish here. So for example, If it makes a difference to have two different products colors shown up in an A B test. Cool. Blue may perform better than red, but did it actually drive more transactions?
Because that is what matters. Not that it drove more clicks in the first place necessarily, but did it actually create lift later on down the road? So those are the kinds of variables that you can start doing on the front end side, right, in terms of the personalization piece. But then the segmentation, the same logic applies.
Does it matter that you have segmentation down to the zip code? level or can it be by the state level and perhaps save you a lot of money on the media front while still driving more or less the same outcome. So it's really looking at these performance based outcomes in the tests but against revenue business goals.
I have to say I think historically marketing has Sort of been a little bit too navel gazy on like the personalization worked. It's like, well, yes, you got better clicks, but in the broader scheme of the lifecycle or the journey that we're actually working on, it's hard to point to any success.
So leveling this up, and this is where the consumer data can play such an important role, actually can show against those bigger business goals and all of your tests and experiments should be run against those, which may take longer and be more complex, I admit, but will actually prove to you that your investments are doing the right things.
Richie Cotton: That sort of sounds obvious when you say that you should be doing tests against, like, actual business outcomes, but I think it does need to be, it does bear repeating, because I'm sure there are so many, like, pointless experiments being run that Yeah people do need to think about this stuff.
Cory Munchbach: No, no fault of anyone's like no one is trying to run things that are not productive, but this is where you either don't have the visibility into it, or, it's not clear. That's like a top down remit. Also, the limitation of the data may, this is the best that you can do. But yeah, I think there's a lot of unfortunate.
kind of barriers to some of that experimentation working as well as it, quite frankly, could. The tools are there. It's, it's oftentimes more the people process side of things where it it runs into issues.
Richie Cotton: Okay, brilliant. So that was experimentation. The other thing we were going to talk about was around data quality. So can you maybe talk through what some sort of common problems are with data quality with respect to customer data?
Cory Munchbach: Truly, there are so many. Common is, is hard. The fact of it is what is common. That there are issues with data quality. there are so many
inherent Tensions in the data space, I
would say, and where you want
to move quickly and therefore perhaps slightly less complete data. you inevitably are trading that off for
maybe higher quality or more complete data, but it takes too long. So you all of these trade offs in terms of complexity of the data do you have a system that can query it? There's all of these variables that lead to Different pieces of the data stack fragmenting out and not necessarily talking to each other in the way that they probably should, but also in a way that leads to problems and data quality.
The other thing about it is just what does it mean to have quality data? There's such a different standard that should be applied. For data that needs to be audited, for example, right? Or, you know, is going to be used for higher level governance types of use cases. Versus if you perhaps get the next best action recommendation wrong on a retailer website, there's no harm, no foul there.
So there's all of these different kind of levers and dials that we essentially need to take into consideration when we're talking about. Data quality and why we need it and how we need it. What I will say is the organizations that do have rigor around their data quality across their systems and clarity here are miles ahead of those who do not, if for no other reason than they are able to move more quickly in implementing the types of marketing and sort of downstream use cases that we've talked about.
It's so frustrating when a customer wants to get started, has great ideas for use cases, only to realize that for years the ERP system or the data warehouse that they've had very little access to has actually been tracking data incorrectly for a long time, right? they're sort of subjected to Type of situation.
I mean, that's frustrating for everybody. And so I don't necessarily have there's no magic wands here for sure. And I don't think necessarily every company is identical. But I guess I will say that the data quality question is one that warrants. Much more attention than it gets. And I think marketers in particular would do well by themselves to care more about that.
They want to move quickly. understand that entirely, but they often can end up kind of creating a silo that comes back to bite them through the rest of the organization. If more time were spent up front on the data quality piece, it would actually go a long way downstream for them, but also save them some kind of institutional hurdles that inevitably arise.
Once their new systems and their ways of using that data come to light.
Richie Cotton: I do like the idea that you should probably pay more attention to data that's going to be audited compared to just like, oh, what's the next. Recommendation to click on? Um, yeah, okay. So I'd like to take a little uh, sidestep here into something I know you've been an advocate of, which is data clean rooms.
So, can you just tell me a bit about what is a data clean room and who needs one?
Cory Munchbach: Sure. Like, I think you should ask this question of all of your guests because you're probably going to get a whole bunch of different answers from all of us. At its most basic, it is a mechanism for multiple entities, at least two if not more, to put data into in sort of a sandbox type of environment.
And in a privacy compliant way match those data assets up. So one of a very common way to think about this is for advertisers. That's like one of the biggest drivers here where they want to match up their data against the segments from a streaming platform, for example, or a DSP and it's a way to say, okay.
We've got these 10, 000 people, we want to go find them. We can't see that data exactly, but we trust that you'll take it and then go target against those initial 10, 000 people that we put in. The purpose of this is because in the deprecation of the cookie and in a more privacy compliant world, you just can't go in and play with that data directly.
We need to have more protection for anonymity being able to just overall preserve those assets. What I think is going to be super interesting about these is that you're going to have the opportunity to create kind of more scalable data networks where there's a lot more control from the contributing parties into the clean room, so to pull that data back out.
So you're not necessarily having to co mingle it permanently. You don't have to transform it in ways that won't last forever. The challenge right now is that there are tons of different types of clean rooms. All the big tech platforms have them the streamers and the advertising platforms have them, but each serves a bit of a different use case and it's early technology.
So I think we'll see a lot more standardization of what we can expect through all of these different tools. I expect also that sort of the cost model will catch up with that as everyone figures out. How do we do some of the monetization and activation that we've been doing again in DSPs and other platforms in a not privacy compliant way?
What does that look like in this new world without cookies and somebody's other identifiers to match up? And data clean rooms just are going to play a really critical role in being able to support some of these systems. But doing so with more privacy in mind and baked in really structurally to those types of tools.
Richie Cotton: Okay, so this sounds like it's a pretty clever solution to the idea that you need to be able to target specific customers or specific consumers, but without sharing data from other companies.
Cory Munchbach: That's the idea or at least being able to when you share, you're doing it in a way that is trusted on that you have a lot more control. So whether you're just trying to find new insights, whether it's measurements, look alike audiences, there's a lot of different use cases here. But the key is that two parties are opting into that and can basically opt back out while still protecting the integrity of the data that they put in.
And that is so critical that that lack of transference of ownership yet. Transcribed And kind of the lineage there is really where ThingRooms shine compared to anything we've ever had before.
Richie Cotton: Okay. And you mentioned that there are a few products for this out there. Is this something companies might want to build themselves or is it just like, no, just go and buy something that exists
Cory Munchbach: at this point I would say you're gonna need to buy something that exists because without the integrations with other assets you've essentially just kind of built an empty database and needed to figure out how to do differential cleaning and normalization and pseudonymization and all these kind of key processes but without the benefit of the data coming in.
So at least for the time being. It's very likely that the use cases that you need to implement you can get that through a ad partner or a data partner that has a cleanroom offering that's already connected to the other types of data that you will ultimately need to put into that cleanroom.
Richie Cotton: Because this sounds a bit about separating different data sets into different places. And before you were saying data silos are a terrible thing. So just to make sure that everyone understands, like, how is a clean room different to a data silo?
Cory Munchbach: So your data shouldn't live there. Like, you should think of it much more as a, a transference than it is a place that data goes to live. So once you've done The matching, for example, at an advertising use case, I've put my 1000 IDs in, they've been matched and now the ad platform can take that back within their own data set.
You can essentially retract your 1000 IDs. It's much more about a kind of place to come together and say, Okay, here's what you need to know. Now go find those people in your world. Whereas before it would be here. You have to actually give all that data over. And now the ad platform has it and they've taken possession of it.
So it's. That and even within your own ecosystem, it's not supposed to live there. It's supposed to be a repository, a temporary repository for that sort of transfer of insight to go or that process to happen. And then the data should be ultimately removed again.
Richie Cotton: And have you, I know you said it's early days, but have you seen any success stories from these clean rooms being used already?
Cory Munchbach: Yeah, I mean, customers are so innovative about how they're thinking about this. I think one of the things that I'm most excited about is actually being able to, I alluded to this before, kind of with the second party. Data, which is that you have a CPG and a retailer coming together and exchanging data from a loyalty program an advertising program.
For example, like how did these things actually work? Another example would be now that again, cookies going away on websites. You can't have a referral link anymore because that's going to be seen as a third party cookie in advertising. On another website. So how do you as the advertiser know if someone clicked?
You're going to need to get that data directly from the place the ad showed up. And so clean rooms can be a place where The commerce platform tells you who purchased and you can say, okay, these are the people that I targeted and actually get a reasonably good measurement for that without having to directly refer them back.
So it's a lot of these new ways of doing old things in the absence of the kind of compliance of the ways that we did them before.
Richie Cotton: Okay. So, uh, obeying all the new privacy laws, still getting the same benefits of sort of finding out about the the consumers.
Cory Munchbach: Certainly different benefits, right? It's going to be it's not going to look exactly the same. And I think that's where we still see so much experimentation on this front as people are still figuring out what that looks like. But yeah, we're, we're in this process of transferring one of the biggest shifts probably in the history of advertising of how we do things.
and clean rooms are going to be a crucial piece of being able to do that. And still having the kind of control and insight that you need to, to justify the spend. So yeah, they're critical, even if it's still early days.
Richie Cotton: All right. Super. I would like to spend a little bit of time talking about skills as well. So it seems like we've talked about some sort of data. Quality skills, some experimentation skills for people who are interested in customer analytics. Are there any particular skills you think they need to learn?
Cory Munchbach: It's a very technical role. I think technical marketers, technical data engineers are really starting to show up in marketing or more business oriented parts of the organization in ways we haven't seen before. I was just having a conversation with a partner of ours recently who talked about Five years ago, they didn't have any data skills in their marketing organization, and now they have they have data engineers, they have sort of data quality folks, they have folks with expertise, whether it's knowing SQL or other kind of coding languages on the really technical side.
Certainly more data scientists folks who at least have a passing familiarity with Python and things like that that's going to be very much more a valuable addition to a marketing organization. And then on the customer level side. There's a lot of tools that require you to, again, passing knowledge, at least, of HTML, for example, if you're working on the website of things.
So, there's a core skill set that the more you have of it, whether it's on the data science, analytics, or in the harder core types of disciplines like SQL and things like that there's a nearly unquenchable demand for that on the business side as marketing tries to figure out how to bring these things closer and closer together.
And they're behind, of course, right? Those have typically lived in I. T. or in, in agencies before bringing those kinds of things in house and being familiar with all the right tools that you need, the analytics platforms, certainly C. D. P. Web tools, like those are the core competencies of a, of a good Martech stack and having knowledge of how those work and, and being able to sort of have.
At least like entry level marketing ops kind of chops will set apart candidates now. And again, still seeing wholly new roles coming together that I think we're going to see a lot of really cool, interesting things coming out of the next few years as these new skills kind of get blended into these unique new types of marketing roles.
Richie Cotton: Excellent. And it just seemed the all these technical skills is very different from like the Mad Men TV show is all sort of drinking cocktails at lunch and coming up with ideas. So, definitely a change in the field.
Cory Munchbach: Very much, very much so. But I also think Like, I, I don't want to get away too far from the idea that the real unicorns in this space, you know, for the folks who are really thinking about how they set themselves apart is, they really intimately understand their business and their consumers as well. So it's not just enough to have the skills it's also being able to blend those three, kind of the technical, the business, and the consumer.
together. That is a super potent combination. And the folks who can, put those different hats on and communicate with those different parts of the organization, like those are really going to be the unicorns of, of the next few years for sure.
Richie Cotton: Fantastic. All right. So what are you most excited about in marketing analytics and customer analytics at the moment?
Cory Munchbach: Yeah, we talked about one thing. I think that the clean room rise and sort of starting to see that take off is super interesting and exciting to me of like where that can take us I guess, excited, nervous. We'll see. We're well on track, supposedly for the cookie to really start making its full on exit in January of 2024 per Google.
So I think that's just going to create it. Space for like a lot of really new interesting ways to measure and understand how consumers are engaging with you. It's going to be a chaotic mess, but from that usually comes some pretty exciting innovation. So that excites me a lot. I also really do.
I can't stress enough that I think that privacy as a competitive differentiator is really something to watch. And I'm excited about the brands that I see who are baking that in from scratch and really making that a part of what they do. I think that's going to be pretty remarkable to watch and to see how that unfolds.
So in all of those areas, all touched by data and how you use it and how you think about the consumer. I think all of that is incredibly exciting and kind of on the verge of some, some big innovation or disruption over the next few years.
Richie Cotton: I love the idea that caring about customer privacy is now a competitive advantage. In a way that maybe it wasn't a decade ago. That's brilliant. Okay. Do you have any final advice for any organizations wanting to improve their customer experience using data?
Cory Munchbach: so I think I've said it a few times. This for me is all about kind of creating that institutional muscle. So maybe the, the single piece of advice is if it's a type of organization that is prone to thinking huge and only being able to do the really big stuff think smaller, think in kind of fragment bite size progress that you can take a step toward every single day.
I feel the companies that are really showing the most progress in this area have taken that kind of bite sized or fragment based approach and building, building up toward it. And so that would be the advice is making sure that that's how you're thinking, how you're approaching and trust that over the same period of time, you will actually be a lot further along for having showed up.
and done a little bit every day than just trying to wait for like the one big day where all of this explodes for you. It's that discipline to, to work at it every single day that really makes the big difference.
Richie Cotton: Fantastic. I like that idea. Just inching towards progress. Brilliant. Okay. All right. Thank you very much for joining me on the show, Corey.
Cory Munchbach: Thanks very much for having me.
How to Become a Prompt Engineer: A Comprehensive Guide