Shane McAndrew, Chief Data Strategy Officer at Mindshare joins Allyson and Brett to reflect on what making meaningful connections with data really means.
Join them as they explore how brands can be more human when it comes to their data strategy, and as they examine today’s challenges from a marketing perspective. Listen in for McAndrew’s take on the pros of first party data, media fragmentation, and how we make the next iteration of digital better. Just don’t expect him to put in a good word for cookies.
Allyson Dietz: After a bit of time off, we are back with a No Hype podcast, and with us today we have Shane McAndrew. Shane marries the science of data and analytics with media creativity to accelerate growth from Mindshare clients worldwide. He brings more than 20 years of experience across digital, marketing, and ad technology to his role, with particularly deep e- expertise in addressable media, programmatic technology, and analytics. Leaning on his client service and product background, Shane delivers best-in-class integrated brand and performance services for his clients. Shane, thanks for joining us.
Shane McAndrew: Oh, it's a pleasure to be here. Thanks for having me.
Brett House: Hi. Hi, Shane. Uh, yeah. So you've spoken a lot. You know, we've, we've talked quite a bit in the past about leveraging data to create more meaningful connections, you know, kind of a hype-y term. Um, I know that Mindshare references that as, as sort of precisely human in data strategy and approach. And I wanted to get, uh, your perspective on, on really what that means. If you had to deconstruct that notion of, of being more human or more people-based in your data. Data strategy from a marketing perspective. How would you define that?
SA: Yeah, no. So I appreciate the question and yeah, we are, we are in the business of making more meaningful connections. And I'll, I'll start by kind of breaking that down. I think when we think about data strategy here at Mindshare and the application, the communications, particularly at scale, we think about adding or enhancing two things. The first is relevance, right? Often confused, I think, with personalization, but we're hyper-focused here on relevance. So understanding the context of the experience that the consumer's having, right? The human is having. What we know about that environment, what we know about that person, right? Did we use this as an opportunity to deliver a relevant message with all of that information that we have? And the second is additive, right? Can we use data to either know a little bit more about that person, to feel a little bit closer or close the proximity between the brand and the person?Can we have a little bit more fun, right? Or could we be simply helpful or entertaining? So we want to be relevant and we want to be additive when we're designing the experience as it relates to the application of data.
Now, Brett, you mentioned precisely human, and that's kind of a construct we've created here at Mindshare, which is a, a constant reminder to bring together both the accuracy that data delivers, uh, as well as empathy, which I think is something we're hearing a lot more about these days. And it's something that we're, we're almost maniacally focused on. So how can we use the rational understanding of what we know about you and the behaviors that you're having? And then go a level or two, or three or four deeper and start to really understand the emotion, right? The emotional state you may be in when you're consuming the media, and how the brand may be able to key into that and as I said, be additive or enhance the experience. Um, you know, if, if, if we think that someone's in a particularly good mood, wonderful! Let's continue with the joyous content and deliver a message that's joyous, you know, and kind of supports that, or vice versa. So it's really about bringing together accuracy and empathy and doing that at scale.
BH: Yeah, that's a good point. How would you characterize, um, the difference I think, between relevance and persona- personalization? You've kind of mentioned a couple of these terms. Empathy, empathetic use of, of marketing, right? Relevance, personalization, and what's really powering that behind the scenes once you sort of open up the hood? What's actually doing the predictions to understand what's relevant, what's not, what's personal, what's not, what's empathetic or tied to a particular emotional state or not?
SA: Yeah. So, you know, I think if I understand the question correctly, um, you know, first of all, it's the people that we have on the planning side. Increasingly, the, the, the difference between a planner and an analyst is isn't, you know, we're actually seeing that our planners very much the planner of the future, but the planners that we're increasingly putting at the driver's seat are analysts by nature. They come from our analytics teams and or still are, still are in them. And the tools that we're equipping them with are a little bit broader than the tools that, that I grew up with as an analyst, right? We were looking at log files and scouring ones and zeros to understand patterns, right? And try to exploit those patterns to our client's benefit. And these tended to be demographics, and they tended to be behaviors. And I think that demographics and behaviors are fine, and they've gotten us pretty far. But I think, you know, there are a lot more tools and technology these days to help us understand emotion. Um, both the emotion, um, resident in the content and the words that are on a page that someone is consuming, but also the emotion- the emotional state that someone may be in while they're consuming these things, independent of the actual content. And so we've invested quite heavily in cognitive intelligence products that allow us to bring different types of signals, and that's the precisely human right, what we've depended on – the demographics, some light psychographics as well as behaviors – and we've begun to marry it with a much di- deeper psychographic understanding, and a much deeper cognitive understanding of the experience that someone may be having while they're consuming that media. And together, right? In a very precisely human way, uh, those allow us to deliver empathetic executions, right? Which are our terminology for the experiences that we deliver based on top of that expanded data signal. So we're, we're, we're increasingly one part analyst and one part cognitive psychologist here at Mindshare.
AD: Well, what I hear you describing is sort of the future wave of, of, uh, analytics and measurement, truthfully. You know, you talked about accuracy and it seemed like you were talking about the next wave of something like reach of, you know, who are you reaching, how are you reaching them? What is the precision associated with making sure that you're, you're reaching an audience and, and the audience that you intended to reach. And then combining that with more of these engagement metrics. You know, we've always historically focused on things like viewability, but you're sort of talking about it and describing it as that next stage of more of the engagement with the content. Is that fair? Is that a fair assessment?
SA: Absolutely. What we're, what we're trying to, what we're searching for are, um, previously invisible levers and or, you know, more dimensionality to the experience. And we know it exists, right? We've seen it in panels, we've seen it in small doses. So the, the magic has been how can we actually begin to see these things at scale and in a way that is actionable , right? We've often, as as an industry have a tremendous amount of insight over here and a ton of action over here. And one of the things that, you know, tools like identity and the advent and the scale of identity has done, has allowed companies like ourselves to connect those two. And the better we connect, you know, the insight that we get to action, the more empathetic the execution we're gonna be able to deliver. Now, if you add on to that, um, that in the insights bucket, now we're increasingly, um, measuring things like the emotional response and resonance to creative and to media, and the combination thereof. Now imagine that we've got a construct, or we can harness that and push that back in to each experience that we deliver to a consumer, progressively over time. And that's, that's really what we're talking about here. So, yes, it's, it's kind of intent, attention plus engagement. Allyson: Mm-hmm.> equals a new score that we're looking at to help us measure the efficacy of ads.
BH: So you sort, sort of talked about consumer insights on one side. So think of that as your data and all of the data that you have, which is, a lot. And then the actionability of that data on the other side, what's in the middle? Is it ai? Is it machine learning? Is it predictive analytics capabilities that are–
SA: It's all of those, right? It's all of those things. We thought we were gonna get the easy button, much like what we thought with multi-touch attribution, uh, you know, 5, 6, 7 years ago when, when identity was prevalent. Now that's, that's only in, in, in North America, right? When we're talking globally, of course, we don't have these identity constructs in the data to support it. So we knew from a global perspective that was never gonna be the single solution. But in North America, um, where I've spent a lot of my time, if not most of my time, um, identity was a wonderful conduit or bridge to connect those two things. Um, but now we're seeing identity, right? Starting to, I don't, wouldn't say it's under attack. It's got a wonderful role, but I think the role outside of permissioned or consented records these days, of course, is gonna limit the scale. So it's not identity for us, it's, it's a, it's a, it's a wonderful understanding of the data signal we can get to understand who a person is. Identity, um, married with machine learning, we very much believe that identity in the future is a data science problem.
AD: So what you're des– I mean, effectively that's more, you're looking at more data points, more access to data, not less access to data. Cause we tend to talk a lot about data deprecation, and what you're describing is, you know, different sources of data. Harmonizing those data sources into a single response model to understand what's working and what's not working.
SA: Well, 100%. And I think, you know, what's incredibly exciting, and we're on the lessons that we learned from the days of CRM and relationship marketing and those types of data sets,um, is that today we do still have enough data re– respect within the, within the bounds of ethics and respect and privacy, which I'm, I, I, suspect we'll talk about a little bit later. There is absolutely new and unique sets of data that are becoming available we've not had access to before. And when you marry that with enhanced data storage and data processing capabilities and breakthroughs on machine learning, um, we're able to apply these at speed and at scale to make decisions in places and spaces we never have before. At Mindshare, for example, we've invested heavily in automating a lot of this process so we can make it very affordable and cost effective for our clients. We've scaled auto machine learning with a partnership with a company DataRobot back in my hometown of Boston, to be able to push, you know, machine learning up in– way into the planning phases. Um, incredibly exciting.
AD: Yeah. So you started talking a lot about more data and, you know, permissible uses of data. I think one topic that comes up a lot in the industry is the, the interest and the growth and the focus, particularly around advertisers, around building out first-party data, and building up your own first-party data. You know, what do you see as the pros and the cons of first party data for an advertiser?
SA: Yeah, so I think a couple of things, if I could back up almost a little bit. It's, it's, you, we've touched upon quality quite a bit, right? I, I know that we'll move into the world of quality, particularly as we move into a world of machine learning and artificial intelligence. I believe very strongly, I would rather have small, stable, accurate data sets than large sparse inaccurate data sets, which is what plagued us for a very long time. And I'll connect that back to the cookie. Um, and my desire for it to go away as quickly as possible later in the conversation, I'm sure , um, it cannot happen fast enough. As a matter of fact, we act today like it already has. But, um, I've taken myself way off track. What was that question? Oh, the pros and cons of first party data. Yes. Um,
AD: So it's ok, you can throw the cookie under the bus in the meantime. I—
SA: I'm gonna take every opportunity, Allyson, that I get to do that. Um… Listen, there are a lot of pros of first party data and, and a lot of our clients are in a veritable arms race to amass as much first party data as they can. Now, we caution them, right? To move down that road with care, um, and with great consideration. We are pro-first-party data, myriad benefits. Uh, the biggest ones we see are often if done right, it's less expensive. Right? Often, again, if done right, it's less biased, um, it's less available to your competition. Right? And it's, it's more easily available to you. So all of those things come together and they create a really awesome, um, tool for our clients. Absolutely. Absolutely. Representative of the clients they have today. Yeah. Right. But there are some cons, right? There are some cons to that, which I think you asked as well, which is it's less scaled, right? And the bigger the, the lower the likelihood or the bigger the brand's total addressable marketplace or TAM, the lower the likelihood they're ever gonna have, more than a small percentage of their potential customers. So again, this goes back to it's not about having it all, it's about having the right sets of data and the right analytics or, or, or data science tools to extrapolate and predict, make predictions off those smaller data sets into your prospecting or unknown populations, right? But therein lies and other con: are your customers of today, representative of your growth opportunities of tomorrow? And I would say sometimes, right? But you, you, you, I, I just am saying all of this to caution clients from ever falling back on that as a single data strate-, a single pillar of data strategy.
BH: Yeah. Yeah that, that lookalike modeling will solve, like lookalike modeling all of my prospecting load . Cause I know from limited data, exactly what attributes are driving, uh, uh, customer acquisition. So I'm gonna just double down on that from now until the end of time. And and you're saying that's also highly limited from an approach perspective.
SA: I think that's highly limited, and at, and at this juncture, I think it should play a very small role in a portfolio. But soon I'll be talking about lookalike modeling the same way we're talking about cookies, or at least I am, right? I think it, I think it's an easy button. I think there's some value there. Um, I don't think that the customers that are collected, however they're collected or representative of the total addressable market that a client has, there's some polls there, um, nor growth or, or new segments. So I think the first party data has, has its place, it's incredibly valuable, but it needs to be one part of a balanced portfolio in your data strategy.
AD: Yeah. It's interesting that you, I have a question that, you know, I'm curious, I'm genuinely curious about your, your answer here, which is, if, if you don't, if you think that there, you talked a lot about engagement and sort of driving the brand side of things, and then also the precision and the accuracy. If we feel like first party data is a portion of the strategy, but not the whole strategy, what be, what makes up the rest of the strategy? What, what are the other components and what are the other things that brands should be considering in, in that world where, you know, data deprivation does exist and there and it's, you know, less Easy to reach the intended audience?
SA: So I think dimensionalizing, well, first of all, when I, when I advise clients on first party data, it always starts with how that data's collected. What we wanna do with their first party data is make sure that the data is as representative of the population they seek to serve, right? As they intended to be. And that's not always the case. And that's not, that's never intentional, right? From brands, the easy thing is to say, okay, look at the brands that are highly promotional and are throwing out t-shirts and ball caps and anything else just to get an email address. Absolutely not the right way to go about this. We've learned that over and over for decades in the world of CRM in relationship marketing. But a lot of that promotional activity still exists as a mechanism or masquerading as a value exchange, right? I would say that that's,
AD: We only need so many peewee hermanns @gmail.com in your, in your CRM database.
SA: Right? You only need so many of those. That's, that's exactly right. But the other thing I would say is, it's not just what's the, what's the value exchange, but where and how are you collecting this data, right? If you're, if you're a brand that primarily distributes offline and you're collecting these via online panels or through a wonderful value exchange, whatever that may be, how many of your customers are actually there and, and is that representative of them? And is that scaled? So, we kind of have to look at the sales data to see who's buying and who's purchasing, and we constantly marry that up against who we're acquiring in terms of consented records. And saying, how, how, how are these, do these look similar and are they off or are they not? And that provides a good guidepost for us to understand how representative the consented records we're getting are of the people who are buying your products. Not to say the future buyers of your products, but the ones buying today. Again, it's just one thing we look at to help brands understand and calibrate: are the data collection mechanisms as representative of the people they seek to service as they could be.
BH: Yeah. And, and this sounds like you're tying into that theme of sort of ethical use of data, but also the, the theme of which I think lends itself to this notion of consent and where consent is captured, what consumers are consenting to, and whether or not, not they know, uh, precisely what they're consenting to. Because if their data's being used in ways that they weren't aware of, or they didn't read the fine print, you know, that's gonna affect the experience downstream negatively. So how, how do brands manage that? Because the complexities around how you collect data, where you collect data and what you collect, and then also what consumer sentiment is on that other side of the fence saying, I understand how my data's being used. How do you help clients bridge that?
SA: Huge gap there, in my opinion, and a huge opportunity, right? And we've, we've focused a lot of, uh, on that gap here, um, at Mindshare and group M um, effectively the roadmap we see is you've got data collection, data hygiene, data governance, and we believe we've added a fourth layer that we hope more people focus on, which is data ethics. So in our data governance model, that's the fourth layer that we focus on quite a bit because we believe there's a massive space between consumer awareness, which is growing, and sentiment, which is not growing, or it's actually increasingly negative and regulation or legislation and actually ethics, right? If we look at the regulation, it's bare minimum if you put yourself in the center of the decisions that are being made, um, we don't think that there is, you know, if, if I, as a consumer, I don't believe that the regulations that are happening now, um, protect me from an ethics standpoint. It's simply all, all a brand needs to do is have my consent. And I probably don't understand, to your point, Brett, the consent and what it means in terms of the uses of the data. Probably never will. Maybe you don't have to. And that's where the brand comes in. And the opportunity for a brand, I think, to engender love and loyalty from new customers and, and all the, like, by actually going beyond legislation and thinking about ethics, we often say just because you can doesn't mean you should. And that goes for data being collected, data being stored, and data being applied. Um,
BH: Yeah, and and does that application also mean communication to, to, to your customers? Especially if you have things like email address. Let them know how you're using their data, how you intend to use their data?
BH: And the val– and, and really communicating a value exchange to what you said earlier.
SA: Absolutely! And I think that's part of the value exchange and open, honest communication and dialogue, and it's a two-way dialogue. And I think this is what's really exciting about media and the new formats that are coming out that are more dynamic, they're more interactive, hell they're shoppable now, and the line that's constantly moving between paid own and earned or CRM and owned assets and media, right? That line is absolutely blurring. And what that means to us, and, and the reason personally that I've left CRM companies and come to media company is, wow, now I can apply data in all of the fascinating ways that I could with direct communications, but I can do it at a grand scale and I can do it in a digital format that's increasingly interactive, right? And now we're moving into TV with CTV as digital and addressability permeate everything, right? So now things get really, really exciting. The media world, which was typically a one way communication vehicle, is now bidirectional. And the ad units are incredibly exciting and incredibly dynamic and created– creativity, I believe is, is coming back and going through the roof and recoupling media and creative and all of these things lead to an instrument that allows a better dialogue. And that better dialogue will lead to an emotional connection, which will lead to, uh, to a relationship. And we do believe that that's possible, and we're seeing it happen and we're proving it out.
AD: So, I mean, effectively you can attack the first party data strategy more so through your, your ad and, and your medium strategy, essentially. You can build that first party data more so through that engagement and through that shopability and create more customer engagement that way, is what you’re suggesting.
SA: Absolutely. I mean, just think about the scale and it's, we've gone so far as to create a process we call Data by Design. And I mentioned before we're seeing more analysts sitting in planner's seats. And you know, that's led to a shift in the thinking. It used to be, very recently, what data is available and what's my data strategy to drive my communications, to drive that relevance, to drive results in business outcomes for my clients. Now we're pairing that with a process I mentioned called Data by Design, which is planning data outputs, everything from consent to other artifacts or pieces of data that we would be able to use ethically to deliver an additive and progressive experience. The next time that we come across that consumer, or dare I say, a consumer liked them .
BH: Yeah. And, and, and that ties back to that notion of actionability. You've got consumer insights on one side. You've got this machine learning layer that's doing predictive, uh, analytics, so to speak, uh, in the middle. And then you've got this actionability side, which is, you know, can come in a lot of different formats. But what I'd like to kind of dive into, cuz I've heard you talked about this a lot, is this notion of, of creative customization, personalization, relevance, however you wanna, however we're gonna put that. How hard is that actually to do? Cause I mean, I've been in the industry for a long time, the versioning process, right? Because your segmentation schema, your, your audience creation schema can be so broad and so deep. How do can a brand actually manage that? And how many clients do you see actually doing this effectively?
SA: Uh….It's hard.
AD: IWas gonna say effectively is the key word there. I feel like Shane, because I think,
SA: Yeah, yeah. I mean there's, how do you
AD: You're not A/B testing in that environment, right?
SA: And how do you do it in a way that doesn't just create clutter? Right? Um, we've got a rich intelligence, we've got rich segmentation that sits across marketing and, and, and cascades if done right into media so that there's consistency across consumer touchpoints. Um, great. We work with marketing and media on all of that. And then you've got the activation and action side. Um, then you've got the creative side, right? And the production units often within the creative agencies, which are pumping out content, often misaligned with the intent of the media plans and the context we're buying, often misaligned from the segmentation and the master segmentation scheme of it's owned by the client. So this is a people problem, right? This is not a technology problem, right? With machine learning, as we've discussed, we can create more versions in a shorter period of time than we've ever been able to. And we can do it based on all the rich signal that supports and drives a segmentation, not a problem. We can actually push it into activation. Not really a problem either. I mean, hell, we see all the ChatGPT conversations going on. Production of content is no longer an issue. I think we actually have to be worried about the opposite, which is too much content and clutter and ads not working well for people and us losing relevance. But that wasn't the question. Is it possible? Right? I think yes. I think it's gonna take, and I think we're seeing this now, a recoupling of media and creative, right? I think that, um, you're gonna see media and creative decoupled so that the intent of the context and the audiences that we buy and all of the intelligence that sits behind them is resident in the shop that's creating all of the versions that support it, measures them and optimizes them so that feedback loop. And I think you're gonna, so that's one jump removed, right? That causes a disconnect. And I think you're seeing media agencies move up the strategic totem pole that they're having conversations with the likes of the Accentures, the Bains, the BCGs, the Deloittes about real strategic segmentation right? In the office of the CMO and even the CFO. So we're understanding the core strategy of the business from a marketing perspective, and we're better able to translate that. So the gap between real marketing strategy and segmentation and the media agency closing, and I think with the recoupling of media and creative, at least on the digital side, um, you're seeing another gap created. So now you've got horizontal process and technologies that can manage that from end to end. And that's not what we've had. So yes, it's entirely possible. We've got all the pieces and the parts. Now we've just gotta organize the people and their processes, um, to accomplish. It's–
BH: The hardest, the hardest part of all.
SA: Hardest part. It's always the people, right? It's, uh, it's the hardest part.
BH: Yeah. And Allyson has seen a lot of this in the front lines of sort of our analytics and attribution, uh, solutions is the notion when you're trying to stand up a client, the, the first challenge we always face is their data, uh, within their data and their silos within their organization. You need data in multiple places. They don't have their house in order, as people like to say. And standing up, uh, an effective analytics practice or an effective data strategy, targeting strategy, requires, you know, that you have a centralized repository of all of your consumer data, that you've got access to it across multiple functions within the org, whether it's tech or marketing or analytics or measurement. Um, and that tends to be the biggest blocker in what takes the most time. And so, um, you know, I go back to that. How practical is it in, are you seeing progress in that?
SA: In that we're, we're seeing a lot of progress and I'm, I'm so enthused by the progress we've been seeing in the last 18 months, particularly on the connection of, um, strategy to media intelligence, to creative intelligence and the connection of those. And actually the measurement loop, right? That roundtrip ticket from the ad exposure back to some sort of customer response to that stimulus, we're just getting better and better. And, and with machine learning, those feedback loops are coming faster and moving faster and delivering more and more iterations of what we hope to be much more relevant and additive content.
AD: So you, you brought up analytics, so let's go there next, since I know you have a background there. And you've spent, now that I know you've spent time, um, dealing with MTA, um, personally, it seems like we all believe that marketing analytics is having to evolve in the new ecosystem, right? And I think that you mentioned earlier there's, you know, there's promises associated with MTA, and I think as the data, um, ecosystem has evolved, that makes it more challenging to stand up attribution. But, you know, how has marketing analytics cha–, have you seen marketing analytics change for Mindshare and for your clients?
SA: Yeah, yeah, absolutely. So, um, it's evolving really, really quickly, right? As, as we've kind of discussed, I think that now that we've moved into a world where we can automate a lot of the machine learning operations to take care of the data organization and data data hygiene and apply automated machine learning on top of it, we're able to efficiently apply analytics to more and more of the decisions that are made across the media workflow. And of course, the hypothesis there is the more data and analytics that we apply to, the more dis– to more decisions, the better the outcomes should be. And that's a thesis that we continually, that is obvious, but we continue to prove out day in and day out. Now, measurement, right? The measurement portion of the larger analytics world, um, has changed incredibly rapidly. I think that we've become too siloed, data's become too disparate, um, and we're just not able to recreate a consumer journey across touchpoints in time series order and then apply the math to it. We're just, we're just not gonna be able to do that. I see the industry moving.
AD: What limits that–
SA: a little bit of, sorry?
AD: What do you think limits that? What, what, what are the limitations that, that we have to, that we're facing?
SA: It's a data limitation. It's not a math limitation, right? It's, it's, it's, the data has become too siloed and the currency of the data has become too differentiated , right? Whether you're talking about aggregate or individual level, whether you're talking about, you know, the actual way that the data is measured or whether you're talking about walled gardens that we all know, you know, are, are keeping that data for themselves. Um, there's just too many holes, in my opinion, for a large online and offline full funnel marketer to employ as a standalone.
As a standalone, yeah. I think that, you know, there is not gonna be a single silver bullet system that allows, uh, accurate measurement. I think in our world, there's kind of three legs to the proverbial measurement stool here. We've got, uh, econometrics, right? We do a lot of offline media and econometrics are still critical to that. , and they have been to unified, you know, measurement systems across the board. So that's a mainstay. We're doing more of that than we ever have.
BH: And, and can you, can you define that, uh, you know, I know what it is, but just for the audience, can you define econometrics and how you're leveraging?
SA: Uh, media mix modeling? What we're looking at is a time series of exposure and response, typically sales data over long periods of time. And we're finding correlation and causation between stimulus and response. In it, in its simplest terms, at an aggregate level. Now we know that, that, that there is some accuracy and there's strategic value to this, but it's not as fast, um, it's not as granular and it's often not as cheap as marketers would like. And all of those have, you know, have been barriers for, for the 25 years that I've been building and now, um, delivering these solutions to my clients. I think that, you know, what's exciting now is software that allows us and, uh, analytic horsepower that allows us to execute more multivariate tests at a greater speed and greater scale than we have before. And what I'm describing is incrementality testing. . So I think that, you know, what we're looking at right now is one part econometrics, one part increme– incrementality testing, and one part channel contribution, which is a slight shift from attribution. We're really looking at, through the lens of incrementality, what's the relative contribution of each channel, and what are the synergies between those channels as it relates to the contribution. So we've moved kind of from attribution to contribution in terms of the language that we used, and those are the three techniques that we're using, um, and pairing differently for different objectives. But those are the three general techniques that we've invested in as it relates to measurement. Now the world of predictions is entirely different, but, um, that's how we're helping our clients understand if their, if their investments have, have paid off.
AD: And are you then, and, so you talked a little bit about different things and you, one thing in particular that I wanted to kind of just ask about was around media fragmentation and the challenges there. As new cha– channels emerge, you know, new re things like retail media networks and new streaming platforms, you know, how, how should brands approach those new, those new channels in terms of measuring incrementality and, and capturing that across the, across their measurement platform?
SA: Yeah, it, it's, and I hate to say it cause it sounds like I said, they're on those, incrementality tests, right? I think now more than ever, brands should be investing in staff who can put together multivariate tests. And the trick is, is not designing the experie– experiment, but pushing the experiment into activation and maintaining the integrity of that experiment and then doing it over and over and over again. And I don't know if I'm answering your question, but I think at the heart of everything we do has to be constant experimentation. And that's one of the things that I've seen, quite honestly, um, fall down, um, in my experience here at Mindshare and other places. And this comes back to a people problem.
AD: Yeah, I was gonna say, it sounds a lot like the creative, testing we were talking about earlier in the, in the, in the flexibility,
SA: I think you have to have a bit of fortitude in, you have to have a long tenure in your seat in order to, to, you know, to kind of drive a learning agenda that aligns with your measurement methodologies that'll gives you the time to set and run the experiments that you need. Um, it's not as if we can run a single experiment. We're running experiments in every campaign that we launch, um, to get this. So I think that rigor and fortitude on the, on the marketing and brand managers, um, over a period of time – years – is what it really takes to fine tune and understand what's working and what's not.
AD: And to be okay with, oh, sorry. I was just gonna say, just, and to also be okay with some of the things that the trade offs you're making, you know, particularly around incrementality testing, if you're gonna hold out a sample, you have to be kind of aware and and comfortable with what you might be leaving on the table there.
SA: That's right. There's sales loss, there's always, there's always sales risk and sales loss for any holdouts or any type of experiment that would, that would, that would lead to a non-treated group.
BH: Yeah. And when, and when talking about some of the organizational challenges, the people problems within organizations, when, when I think of, of those two categories that you just brought up, channel contribution and incrementality testing and channel contribution. There seems to often be a fight for budget, right? I gonna maintain my budget. And so I'm not gonna lie, but I'm gonna put some flavor in my presentations as to how my search budget's performing, how my CTV or linear TV budget is performing to show that I deserve to maintain this budget or, or, or increase it, right. Versus incrementality testing, which I think if done correctly, that sort of multivariate approach can shed some light on really what is driving above baseline factors above sort of endogenous external factors. What's re what marketing is really impacting this final steps to conversion or otherwise?
SA: I couldn't agree more. And you know, the good news is we're seeing a lot more movement on our client's side to drive and enable budget fluidity, which I think has been the downfall of a lot of, a lot of measurement systems. And it goes back to natural human behavior and incentives. Am I incentive financially or ego wise to have a larger budget last year than I did this year and larger than my peers, right? Those systems are the ones from a measurement perspective that are failing. They're stuck, right? So a lot of the consulting that we do within my organization, and I suspect other analytic organizations, is on the org design and incentive alignment side of things. Right? We're seeing a lot of our CMOs get very progressive and, and, and, and move even to zero base budgeting on a quarterly basis as it relates to channels. But it becomes a chicken and the egg, you know, you have to truly trust your measurement system to move in that direction, and sometimes you can't get to a place where you truly trust your measurement system until you make those changes. And so we, you know, you, kinda get stuck and–
BH: I think it takes months and months and months to actually do the planning from a zero-based budgeting perspective, right?
AD: Yeah. Well, it's not just that, it's just, you know, to your point, like if you're making decisions based off of PR analytics, then it becomes really challenging to feel confident and to, and to convince the organization that that's the right choice to be making if you're not already in a position where, you know, you've kind of eliminated some of those human challenges.
SA: Yeah. And it gets more complex as we move from the world of measurement and, and leveraging what history has shown us to make decisions and moving into the world of predictive analytics and getting an organization to employ the scientific method where you're quite literally predicting, executing, validating, rinse, wash, repeat. Sounds very simple. Very hard to move a massive organization the size that we have as clients at Mindshare, we're very lucky to have, but it's hard to move, you know, these larger organizations into that, into that rhythm.
AD: It comes down to trust.
SA: Yeah, it's trust. It's trust, it's all the people thing. It's a human thing. I'll keep, uh, people will get sick of me saying that, but it really comes down to that and, and making sure that you have a right partner and you have advisors that, that you absolutely trust. Um,
AD: I agree. So I think, I think we're talking about trust, and I think one of the things that, you know, we've talked, touched on is, you know, data quality and, and data ethics and, and you know, the, the role that quality and ethics have, I guess I'm kind of curious is how do we make the next iteration of digital better? Well, we keep in mind keeping in mind all the things that we've talked about around data and, um, Empathy.
SA: Yeah. You know, I think it has to be absolutely human-centric. And we know this, it can't, we can't get caught up on this privacy centricity. That's just not human, right? We too often get caught in these ones and zeros and you disassociate what you're doing day in and day out, hands on keyboard, with what a consumer is seeing, even though you're that same person when you punch the clock at the end of the day and go home and start to consume media, right? So I think human-centric design is something that's here to stay and something that, you know, we've embraced wholy and I hope to see others embrace, making sure that the human, um, is at the forefront of every decision that we make. Making sure that you take advantage to build your brand of that gap that sits between privacy, regulation and ethics, and ensuring that your customers today and the customers of the future know that and understand that, um, you are taking a stand. And I think brands have the opportunity to stake it, take the stand on behalf of their consumers, and engender trust and loyalty and build their businesses on that. So it has to be human-centric. Um, it has to be, of course, as a part of that respectful and privacy, safe and permission-based. Um, we have to get out of this surveillance culture and rebuild trust with consumers, and we have to deliver much more compelling and exciting advertising, if I'm being honest. And I think some of the trends that we're seeing, the recoupling of creative and media, better, more balanced signals that are rational and emotional in nature, fueling the decisions as to which media and context we buy, and pairing them with an understanding of the audience, the context and the, and the emotion to deliver an empathetic execution or creative. That's what's gonna get us there. And we're just at the very beginnings of all of this, in my opinion.
BH: I could ask another question, but I think that's, I think that's a good place to end.
Allyson Dietz: I do too.
BH: I I think you just summed it up, you know, and I loved, I loved the topic, uh, this clear distinction between privacy by design versus human, or data by design, and how you've made a clear bifurcation that privacy by design might swing too far in the other direction, right?
SA: They’re two very different places.
AD: It's really about the objectives and the, and to your point, it all starts with the strategy of really understanding what are you trying to achieve as a brand and as an organization, and what does that look like? And then effectively designing what data you need in order to execute that strategy. That's, that's exactly what I heard and thought that was so incredible. Right.
BH: And who are you as a brand, right?
SA: It, absolutely. It's the brand essence. And then, you know, and you're, and then designing an experience and working backwards. I think it's too often that we look around and say, how much data can we amass now let's go figure out what we can do with it. ,
AD: Yeah. Well, Shane, it was a pleasure to talk with you today. We really enjoyed it. Um, and thank you again for joining, for speaking with us, and, and hopefully we can chat again soon.
SA: Yeah, I would love that. It was, it was, it was fun. My pleasure.
BH: Thanks Shane.