This panel discussion took place at February's iMedia Brand Summit in Florida. Doug Weaver, President, Upstream Group, moderated. Panelists were Gerard Broussard, Senior Partner and Director of Media Analytics for MOne, and John Nardone, Executive Vice President - Product Development and Marketing at Marketing Management Analytics (MMA).
Yesterday, the group defined Marketing Mix Modeling. Here's more:
Weaver: I want to start off by stating one of the fundamental push-backs that marketing mix modeling gets and I want you both to comment on this -- maybe John, starting with you. Obviously to statistically model this kind of activity, there’s got to be a lot of money in the pipeline. In other words, you’d have to be spending significantly on the media in order to really get that feedback. Internet and other emerging media, aren’t they at a substantial disadvantage in that mode? In other words, couldn’t marketing mix modeling be used as a justification for continuing to spend a lot of money in TV and print and traditional media? Comment on that.
Nardone: Well, there is truth to the fact that if you’re not spending, or if you’re spending a small amount, it is harder statistically to sift through all of the activity to isolate the impact of that small amount of spending.
That being said, there are ways that interactive marketers can plan and aggregate their data to get a better opportunity for internet to pop in the model. And that really starts with aggregating all of your digital activity so that you’re not trying to look separately at online advertising versus search versus your website versus email, but looking at it as a holistic, digital effort in pulling those together. And generally, if you do that you may find that the total amount spent on digital efforts is enough to show in the market.
Weaver: So even though we really try and preach to the marketer to really look at the different aspects of their digital plan, you’re saying pull all that money together and try and model the whole thing in order to potentially move more money over to this side of the ledger?
Narbone: Well, absolutely. Your first order of business as a marketer is to say, "Is my integrated program working?" and "Should I put more money or less money into that?" In terms of then diagnosing which pieces of that integrated program are working you have some other tools that you can use, standard web diagnostics that can help figure out: "Should I put a little bit more into my website?", "Should I put it into online advertising?", "Is search really driving it?" But I think your first order of business is to say, "For the bucket of digital, how much should we spend?"
Weaver: So maybe that’s our first best practice that the marketers walk out of here with -- if they want internet to pop within that modeling process, aggregate it all together.
Gerard, what would you say to that comment that this tends to lock in the status quo and favor traditional media?
Broussard: There could be, I guess, a self-perpetuated situation in which the heavy spenders show up more than the low budget media. And it's not really just indigenous to the web. We’ve seen cases where radio for example doesn't pop, or newspapers. And there are a number of reasons why that might happen. One could be -- I'm going to throw a word out here: co-linearity. And that means when, let's say the patterns of a medium and its effect towards sales hides behind other factors. So it moves in the same direction as let's say, as TV or print -- let’s say radio does. And you can't really get the effect. You can't see the effect of radio because it moves in the same line as the other media. So…
Weaver: So there may be some implications there for how marketers flight their media …
Weaver: … and if everything kind of runs in a pack, they're not going to really see what's having the effect.
Broussard: If there's more diversity in the patterns of flighting then it's probably easier to pick up the effect of the media.
Weaver: So what's the best practice then, for a marketer who is using this? Does that mean running different media at different times, allowing for quiet periods in certain media where you allow the radio to run by itself? Or, what do you think?
Broussard: I think the best practice for a marketer is to really follow the marketing objectives and try to use the media in the best way they think will impact sales or awareness.
Now, you run, I guess, into a little bit of a trap here. If you’re trying to isolate the internet effects and you're modifying the scheduling of the internet, or any other medium, just to try to show the effect of it, you might not be true to your marketing and media objectives. So it's kind of a catch 22.
Weaver: Right. You can get good sales effect or you can potentially get great research out of it, but sometimes it's hard to get both?
Broussard: That's right. That's right. You have to stay true to your marketing media objectives. There can be some instances when you can isolate, as John eluded to before, maybe a surge in internet spend that would help to quantify its effects.
The other thing that you need to realize is that going into a modeling project, you can't really expect to get effects from all the media. That doesn't mean that they don't exist. In fact you may not even see an effect of a promotion or the seasonality or timing of your advertising. So, you have to take some of the heat off just the internet because it may not be showing up, and say that modeling as a practice is not really a perfect science.
Narbone: I think that's important. Modeling is a decision-support tool, and as a decision-support tool it's not a substitute for judgment and experience, in actually making your decisions. What it does …
Weaver: But sometimes it's probably used that way.
Narbone: Well, sometimes it is used that way. And we actually have to counsel clients sometimes to be careful that they don't have the law of unintended consequences, in terms of stifling innovation where brand managers tend to get very conservative and just stick with things that give them consistently higher ROI scores, rather than experimenting and finding the next new opportunity. And so we do things like help clients carve out "x" percent of their budget specifically to spend in areas that aren't popping in their model, and do those with a very specific learning objective so that they can get quantification of their impact over time.
Weaver: So, John, a lot of the client side people who are here today, at this conference, tend to be the interactive point people so they're fighting for share of budget, share of voice within that overall marketing spend at their companies. They're faced with a Brand Manager who says, "Look, I've looked at the media mix modeling reports and you know, darn it, that freestanding insert I do every Sunday is a great needle in my arm, and the 30-second spots, I know I can make sales pop with those." How do they counter that and how do they fight the good fight in diversifying that process and giving other media a chance?
Narbone: Well, I think there are two things I'd say to the interactive marketers in the room. First, of anybody in your companies, you are the best people in the organization to be the champion of marketing effectiveness planning overall. As internet marketers, we actually are used to operating by the data, and having real time data and very granular data far more than any other people in the organization. So from an existing skill base, and familiarity and comfort with analytics, you guys are probably way up on most of your peers in the organization. I would use that to take the high ground and not focus on selling the internet per se, but rather focus on quantifying marketing effectiveness across the channels. Because I think what you'll find a lot of the times is that you spend more on the internet, not by necessarily proving how effective the internet is, but by showing how ineffective some of the traditional measures are. It's not unusual to do a marketing mix model and find out that you're getting slaughtered in your TV spending, and that TV is not paying out at all, and that that's a bad way to spend your money. That happens more times than I can tell you that it comes back the fact that your TV just isn't working very hard for you. And then the next question is: Well, what do I do with that money?
Weaver: Gerard, you probably see examples of that in working with all the clients you work with at MOne. Talk about that discussion a little bit.
Broussard: I’ve got some extreme examples. One is we have a financial online advertiser who offers a stock brokerage service. Almost 50 percent of their budget is online and it definitely shows up in the effects when we do econometric modeling. So there's one case where you have an unusual circumstance, but an isolated incidence where they know it's working and they keep investing in it. And relative to television it's doing at least as good or a little bit better sometimes.
Then, all the way on the other end, you've got a situation where clients are looking to tip their toes into the water. And they're not necessarily certain what they're going to get on a dollar-per-dollar basis when they invest a dollar in the internet versus TV versus print. This is the question that we get asked most often. And really, when you think about it -- and it goes back to the theme of what we were talking about earlier -- is that gee, if you make investments in traditional media and you could say, if you do what you've always done, you'll get what you've always got. And you're probably reaching the threshold of effectiveness in traditional media where you're maxing out on frequency and effective levels of advertising. That's a big opportunity to bring in another medium to get the people who are lighter viewers of television or don't read the print magazines that are on the schedule, and introduce that into the mix, and show it might be effective on a microcosm level. I know that has been done for years. You have to keep banging that message home, that gee, if I don't show up in an econometric model, there are other ways to show that I as an internet advertiser can be as effective. I think there have been many examples. Yahoo! Direct or Consumer Direct has been one example where they've isolated sales just based on the merit of the internet advertising with some consumer goods companies.
Weaver: I’d like to go a little bit deeper on that point John made earlier about how sometimes it demonstrates that an existing medium is producing diminishing returns. Gerard -- let's go to both of you, but can you go a little bit deeper on that and talk about where TV starts to break down in the plan, because maybe the first step toward opening up budgets is demonstrating where the waste and the inefficiency is. Do you see any of that in …? I mean, how are clients misusing TV and how is it not serving them?
Broussard: Inherent in any medium is this old 80/20 rule where 20 percent of the people are doing most of the viewing, and that shows up in your media schedules. So unless you have a very complex message, you don't need the same person to see the advertisement 20 times. So what we try to do is say to our clients that we need more balance. And I know creative is certainly a consideration. But a client, let's say, that's doing a lot of heavy daytime advertising, I think, in some ways may be throwing some dollars out the window.
Weaver: So, is it fair to say that in trying to go after the relatively light TV viewer, that marketers are just throwing a lot of repetition and redundancy at the schedule, and essentially carpet bombing the heavy viewer?
Broussard: Mostly. If they're not, if they go to selective programming, then they're paying a premium to get to the hard-to-reach male 18 to 34.
Broussard: They're paying a premium in sports, for example.
Narbone: One of the things we see pretty consistently across our client base is a reflection of just that -- for clients who have a male target, the internet seems to pop more consistently than a general female target. So for example, financial services and consumer electronics are two categories where we routinely see the internet popping as equal to, or even more cost effective than, television; where as in packaged goods, it doesn't show up nearly as often or as powerfully as it does in those other areas.
Weaver: John, is it possible also, that that's happening because packaged goods is such a broad industry and such a broad audience? Does that make a difference or is it mostly just the male/female split?
Narbone: Well, I think that it's a couple of things. It's the male/female split. I also think it's the purchase process. If you think about the way consumers buy financial services or consumer electronics, or automotive for that matter, they go to the internet to make their purchase decision -- to support the purchase decision. They don't go to the internet to decide what brand of cheese to buy. So, just logically thinking about how the consumer uses the medium, it makes sense that it's going to be much more powerful in terms of driving an actual sale.
Tomorrow: The $64,000 question: Does the internet work within the marketing mix?