We have French content available. View French site.

Video

Marketing Budgets: Decisions, Data, and More Decisions

Bain Partner Cesar Brea invites Keen CEO Greg Dolan and digital marketing executive Randall Beard to discuss the challenge of managing large marketing budgets in the early 2020s.

Video

Marketing Budgets: Decisions, Data, and More Decisions
en

Chief marketing officers today are at a crossroads, with a foundational decade of research behind and a new set of challenges ahead. Greg Dolan, chief executive officer of Keen, and digital marketing senior executive Randall Beard join Bain Partner Cesar Brea in a conversation about marketing budget trends and the steps every company can take to invest in and optimize its marketing analytics operations.

More

Marketing Analytics Insights

Marketers are increasingly using data and analytics to broaden audience insights, uncover emerging trends, and build brand equity.

Read a transcript of the conversation below.

CESAR BREA: Hello, and welcome everybody. Today I have the pleasure of having two very, very experienced and knowledgeable guests on our conversation to talk about the challenge of spending large marketing budgets in the early 2020s.

With me is Greg Dolan, who is the CEO of Keen which is a next generation marketing mix solution that addresses dynamic adjustment of marketing budgets, and does this in both a predictive but also a prescriptive fashion. And, Greg, I'll ask you in a moment to correct anything I missed there. And then also with us—

GREG DOLAN: That's perfect.

BREA: Good. And then also with us is Randall Beard. Randall is a senior executive who does a lot of board and advisory work in digital marketing, media, data and analytics, as well as product innovation. And we're fortunate to have Randall as a Bain advisor as well. So welcome to you both.

RANDALL BEARD: Thank you.

DOLAN: Good to be here.

BREA: Anything you want to clean up and clarify from those introductions before we get into the discussion?

DOLAN: No, it was perfect. So we're really looking to disrupt traditional marketing analytics through software, and feel like with all the fragmentation that we're seeing across media channels, distribution channels, that an always-on solution, leveraging the best information is necessary to guide decision making on going.

BREA: And Randall, anything from your perspective to clarify further, or?

BEARD: I think you hit it about right. I would just say that my experience is focused primarily on the client side as a past CMO. And then I've done tons of predictive analytics at Nielsen, and then performance marketing at Cardlytics. So I've covered the whole gamut prior to doing this.

BREA: Fantastic, so clearly we have highly qualified folks here to talk about this topic. So why this topic and why today? So when we originally were teeing up this conversation and talking about what was interesting, one of the things we observed is that we seem to be at kind of a best of times, worst of times situation at the moment.

On the one hand, we've got a decade's worth of progress evolving from media mix models through multi-touch attribution, and now more and more increasingly advanced testing capabilities to be able to be in market with thousands of tests and be able to see incrementality. And so there's a long arc of progress that you would see from that story.

At the same time, we have a lot of challenges. For whether you believe it's for privacy or for more parochial interests that different media platforms have, privacy is upon us. And we have everything from the death of the cookie to a number of other examples of how, essentially, our ability to do personalization at scale is getting fuzzed up a little bit.

And so we have these two crosscurrents, essentially, running against each other. In the middle of this is the average CMO who has some reasonably large budget, is looking to shift more of it into more accountable channels, is faced with a blizzard of available options, and where measurement is— we've read about Nielsen getting decertified recently, and so forth— where measurement of whether or not we're having impact and reaching the people that we want to is increasingly a challenge.

And so the topic basically today was, well, OK, so now what? So what do we actually do about this? When you're out there, how are you advising people? What do you see them doing?

But maybe we can start by having you comment on the characterization of those two best of times, worst of times that I described. What do you see? Maybe there's some specific examples that come to mind that illustrate the topic at hand.

DOLAN: Do you want to start? I can give you my perspective.

BREA: OK, go ahead.

DOLAN: So I started my career as a consumer packaged goods marketer in the late '90s. So I had responsibility for these big budgets that CMOs have, and had to allocate to optimize against the financial objectives. So with P&L responsibility, it was all about showing growth in the top line and the bottom line.

So I feel like going back to the late '90s early 2000s when decisions were much easier when you're trying to decide between a couple of different traditional vehicles like print, TV, and radio, there was still more of a focus on strictly historical measurement versus decision making.

And I think that's really what we continue to see in the marketplace. And we've seen pendulum swings throughout, digital coming to fruition on the market and becoming very fragmented, and more complexities being added. Then we have kind of in-store activation, retail promotion, and now retailer media networks all adding more complexity to the table.

And Cesar, to your point, now the data at the most granular levels are becoming less and less available to do any kind of analysis. So we believe strongly at Keen that it is about speed to decision making, given the complexity of the decisions that marketers have to make, and the short timelines around having to understand performance, and then continue to adapt as market conditions change.

So I think there was a bit of a pendulum shift from kind of traditional marketing mix where marketers said, hey, that doesn't work, because it doesn't really capture digital. Hey, we have all this now great, granular personal data that we can leverage through attribution.

Well, that ignored a lot of the traditional vehicles. And now we're finding with all the complexity that we really need a holistic model that can leverage the best information possible in service of a decision. So I think that's really where I'm seeing a move.

And there's more accountability that marketers have to— more complexity and then more accountability around the decisions they're making. So they have to make them quickly and not overanalyze at the most granular level, where there's a lot of false precision, in our opinion.

BEARD: Yeah, so like Greg, Cesar, I grew up in CPG land. I spent the first 17 years of my career at Procter & Gamble and then another eight years at American Express in consumer marketing in a different vertical. And through both of those experiences, I mean, I always believe strongly that marketing exists to build a business and to build brand value.

And that leads you to measurement, and I always believed in measurement. But I always also felt like, to use the words of an old boss of mine, the "empha-sis" was on the wrong "sy-llable." Measurement is important, but what's really important is forward-looking decision making.

What are you going to do differently that's going to grow the brand faster, build more brand value? And there hasn't been enough focus on that. And what really attracted me to Keen is I felt like Keen was using technology to do three things that were really important.

Number one is creating forward-looking simulation capabilities based on marketers real business goals. I need to grow x amount, how do I do that? And they had really good forward-looking decision making, hence the name, Keen Decision Systems, not Keen Market Mixed Modeling.

Two is that the models are always on, they're live. And you can go in every month, every quarter, every whatever period of time, and rerun updates and simulations around what you need to do to continuously improve results.

And then the last thing is, they're also right on the leading edge of do it yourself. They have a very good managed support effort behind that. But it really gives the tools directly to the marketer to be able to interact, and look at the business, and figure out what's best to do, as opposed to waiting for a once a year review delivered via PowerPoint.

And I think all of those capabilities play to where the puck is going in the future in terms of marketing measurement and forward-looking optimization capabilities.

BREA: Yeah, I agree. I think those— one of the reasons why it occurred to me to have this conversation is because I thought that, Greg, what you were doing really was right on a number of trends that we were seeing.

One of the things that I want to ask you about, because you obviously have a number of clients, done a number of implementations and so forth is, if there's a spectrum that runs from the annual MMM review, which is looking back at old data to project what you're going to do six months from now, and so therefore is not only obsolete, but too early.

To the other extreme, a trade desk kind of mentality, where you're literally day by day moving in and out of different media, even algorithmically, without human intervention. As you think about the implementations that you're involved with, how would you describe the cadence that people operate with?

How has it changed the analyst-marketer relationship to be operating in a more dynamic way? And also, frankly, how much of the scope of the marketing investment at this point is that accelerated cadence and process relevant to? I wonder if there's some examples of that that illustrate what you're seeing.

DOLAN: Sure. Honestly, it's still evolving, and the market is still maturing. And one of the areas that we really focused on from a development standpoint was the data integrations coming into the system, realizing that having data in the system and in a repeatable way, able to be updated quickly, ingested quickly, and mapped to our system, would allow for an increased cadence of updates and to be able to guide decision making on a more regular basis.

I could tell you that the industry is still at the very nascent part of that change right from dumping a bunch of data over to a consultant, or even some other kind of system that requires a lot of very deep analysis and granularity, to having it automated in a way that helps to drive optimization and decision making on a regular basis.

That's part of the process that we go through with our clients, because a lot of times we're replacing some of those legacy systems. So we're getting all the data into the system. We're helping them, and we guide a little bit more through that change management process at the beginning.

And over time, as Randall said, they become more and more self-serve as the data comes in, and automated. That then enables more of a regular cadence. So we may start with biannual or quarterly updates to models as we're getting them onboarded, and then as they get comfortable with the information coming in, the quality of the data coming into the system, the model results, the scenario results, we see that cadence become more regular.

So that could be monthly, and a lot of our clients over time move to a monthly cadence. And we see that continuing to progress. Because, really, we have data coming in daily and weekly. We could update models at that cadence, and impact decision making for some of those more, shorter buying cycle type of tactics, particularly in the digital world.

So we see that continuing to progress. I think we're at just the starting point of that. But I think that's where the market's going. Data coming in regularly, there's automation in the modeling, there's automation in the optimization that goes directly into buying systems, and that feedback loop continues in a very automated way, driven algorithmically.

BEARD: Cesar, just to add to that, I think as a board member, one of the things that I've had discussions with Greg about is to think not only about our own Keen platform and the data and the tools that we're providing, the marketers, but also to think about the infrastructure that they have or need to enable these kinds of much quicker cadence decisions, and to be able to optimize and improve the results they're getting.

So if you think about things like budgets, the approval and decision-making process, the talent, there's all these things that are peripheral but are very, very important to a client being able to operate at speed that many clients just don't have in place.

And so I think we have a job to help them define and understand, what are the things that you're going to need to do independent but related to bringing our platform on board to actually fully take advantage of it. I mean, just to take a simple example, in the old days when I worked at P&G, to get like a change in the budget, that might take weeks.

So if I wanted to make a decision to next week go out and spend more money and grow my business faster, well I couldn't do that and leverage the platform without going through the decision making process to change my budget level.

And so there are many of these things that I think, as Greg said, are causing clients to still be in the early innings of being able to actually operate at a really fast cadence, outside the world of things like programmatic, where things are much more automated and algorithmic in nature.

BREA: Yeah, Greg, when you mentioned the word "feedback loop" in your comments, and so one of the things that— when we think about these things, it's not just about a technical implementation. It's about behaviorally rewiring a collection of people to operate in a different way with a different set of objectives.

How have you found it effective to represent the feedback? In other words, for example, we could say, oh, here in a classic MMM sense, here's one brand channel combination producing this ROAS at this point on a response curve. And here's another one over here.

And if you trade the two, that makes you X amount of money. That would be an example. I wonder, how have you represented the feedback in a way that motivates behavior? Or what have you learned about when that works and when it doesn't?

DOLAN: So part of our job, as a classic success organization, is making sure that we're a mission critical tool that's driving their decision making. So you talked about the interaction between the analysts and the marketer. But there's also interaction with sales and finance.

As Randall mentioned before, this is a business decision. In my past life, it was always, hey, marketing mix was run by the insights, and analytics, and marketers. And everyone else questioned the results. What we're enabling is full cross-functional, cross-agency perspective on the decision.

So all those assumptions, any plan is locked in our system that's executed, so that we know, when we get data in, we can actualize against that forecast. So once data comes in, we can actually do a due-to analysis. We have waterfall charts in our application that show actual results versus the estimated results, and then what was specifically driving those differences so we can then optimize going forward.

So that's what really starts to continuously validate and drive the type of change and belief in the system that gives the marketer, the rest of the organization the confidence that they're making the right decisions. And it's actually moving the needle on the business.

BEARD: Yeah, I would really second that. I think it's almost like— I think of it as validation, Cesar. The model predicts, OK, if you do this, you're going to get X. And then you actually go do what the model says and you get fairly close to X, and that's worth a certain amount of money.

And you go through that cycle a few times, and the organization sees, yeah, first of all, the model is fairly good. When we do what it suggests, we get generally what the return is that it says we'll get.

And then people start getting real confidence in that and believing that, yeah, this thing is really generating a lot of incremental value for us. And I think you need to go through a few of those cycles to gain trust and credibility throughout the organization, beyond marketing and analytics.

BREA: Yeah, I agree. I think that notion of repetition and trials here, rewiring behavior is not something that happens in a single trial, that's for sure. Let me ask one more question, if I could, which is, what is your favorite example of somebody who's gotten this right? And what did they really emphasize to actually get value and make this part of their regular practice?

DOLAN: So I'll use the example of Post Consumer Brands. We've been working across their portfolio since 2018. And it was a change management process with them. So we started with one brand. We developed some good case studies around performance, and then we were able to expand to other brands. So there was that iterative approach to gaining the confidence of the organization.

And part of that was also activating some of the recommendations coming out of the system, and then seeing the performance improvement. So going from heavy flighting, which a lot of times that we've seen is very unprofitable, to more always on continuity of support.

That drove real performance both in the top line and the bottom line, and drove— got some eyeballs from very senior levels of the organization as well. They started to embrace that as part of their decision-making process. So there's what I call the hard integration and the softer integration into an organization.

The harder is all the work that we did to integrate all the data into the system, so they can actually update very easily on their own every month, but then also becoming part of that mission-critical system that's guiding their decision making. So every month they're updating, or every quarter they're updating.

And our system becomes the reporting system to senior leadership, but also the recommendations for continued optimization going forward. So integrating as part of their decision cycle, and integrating it part of their business processes, I think, is key.

And also from a confidence standpoint, starting to embrace sometimes against the grain recommendations that are really moving the needle, and we're able to validate those over time.

BEARD: Yeah, Cesar, I would just add, as a board member, I don't get directly involved in the clients. And so it's hard for me to point to a specific example like Posts that Greg just did. But what I will say is, across my career, when I've seen organizations really change and embrace the kind of thing that you're describing, two things are almost always true.

One is, there is a senior leader in marketing who establishes a performance-oriented culture and says, guys, women, men, children, whoever is working for me, it's like, marketing is here to drive the business. And we are going to measure that, and we're going to learn how to better use marketing to drive better business results. And having that cultural change from the top is really critical.

And the second thing is having a CFO on board, a finance organization that really buys in, and believes the marketing analytics, and believes the recommendations and supports them, and makes them part of the way they think about growing the top line and the bottom line of the business. When you have those two things going on in tandem, generally you get the kind of change and the results that we've been talking about here.

BREA: Randall, I'm really glad you raised that. Because I think what we've observed is that you can have the best analysis in the world, but if you don't have the operational flexibility and the buy-in to actually move the money and the follow through on that, then you really are nowhere with this.

And it's so easy, given how marketing budgets have been planned and frankly contracted for in the past, to limit that. In particular with a mixed model platform in this case, you're actually thinking, in many cases, about moving money across channels, not just within publishers in a specific one. So I'm really glad you raised that.

BEARD: I was going to say, CMO-CMO alignment is critical. And to your point, there are now, there are modeling capabilities where you move money between or among brands as well. And that's also really important to the CFO. So, anyway, so, yeah, absolutely. Greg, you were going to say something there.

DOLAN: I was just going to add that, particularly as we've seen over the last 18 months with COVID, being able to react very quickly to market shocks and the aftershocks is key, and having an always-on system where you can introduce COVID effects, and consumer behavior shifts is key to the marketer, and making sure that they're making the right decisions at any point in time.

BREA: Well, I want to thank you both. This is actually a really interesting conversation. It's a topic of a lot of personal interest for me, but not just me, but my colleagues as well. This whole notion of marketing acceleration is a central theme for what we see clients wrestling with and that we're certainly trying to help them with. And I look forward to opportunities to partner with you both on helping them to address those challenges. So, once again, thank you.

BEARD: You're welcome, Cesar.

DOLAN: We appreciate it. Thanks again.

BREA: Definitely. So we'll leave it at that, and look forward to our next conversation.

More

Marketing Analytics Insights

Marketers are increasingly using data and analytics to broaden audience insights, uncover emerging trends, and build brand equity.

Mots clés
Mots clés

Vous souhaitez continuer cette conversation ?

Nous aidons des dirigeants du monde entier à matérialiser des impacts et des résultats pérennes et créateurs de valeur dans leurs organisations.