Commentary
Whose Numbers? The Media Debate
- by June 17, 2002
Media: What are some of the newest and best methods you are using to measure the success of online campaigns for your clients?
Jason Oates, director of media and strategic partnerships, iballs, a subsidiary of Avenue A: We’ve been concentrating our efforts on getting as many attitudinal and behavioral measurements as possible. We’ve been working with a couple of different suppliers to help us measure brand impact or the impact digital marketing has on a specific brand, and getting a good gauge of things like awareness, perceived quality, point of differentiation, loyalty, and purchase intent—which are all direct or indirect attributes of brand equity. And then combining or comparing that data with key conversion data that’s all coming from action tags on our clients’ websites. It just gives us a lot better perspective of return on investment. Until we do look at both the attitudinal and the behavioral, we won’t really have a clear enough perspective on these things.
Tim McHale, chief media officer, Tribal DDB: We’ve launched a CRM initiative, and over the last several months, we’ve pulled together a team of people with strong database management capabilities. Many of our clients are not selling anything online, yet have realized that the Internet poses a great communications tool to begin a dialogue.
Michael Jacobson, senior media planner, NOVO: We’ve partnered with a company called Dynamic Logic that helps quantify the branding impact and ROI of online advertising campaigns from a branding standpoint. They measure different branding attitudinal metrics, such as brand awareness, purchase intent, and message association. And what we’re trying to do now is steer some of our clients—instead of looking at traditional online metrics such as click-through—to looking at things such as brand awareness and purchase intent. If we can bump up purchase intent by a couple of percent, that’s huge from a traditional advertiser’s perspective.
Oates: We are working with Ad Index and also Millward Brown. It always has to tie back to someone’s overall marketing objectives. I see so many campaign objectives out there saying, “We want to increase awareness and increase purchase intent.” And then very few people show any metrics or measures that actually show improvements against those objectives. Doing things like the attitudinal research through Ad Index or Brand Impact through Media Metrix really help. What you can do is, once again, go back to the behavioral and the attitudinal: there are behavioral metrics you can look at that are gauges of increased purchase intent. Say, if you’ve got an online mortgage company, how far do people get down the sales process? If we can get people 75 percent of the way through that application as opposed to 50 percent through, then we’re actually seeing behavioral metrics that show us an increase in intent. Then on top of that, if you layer in the attitudinal data and show a 10 percent increase in purchase intent, then you know you’re going in the right direction.
Media: Do you think that a lot of campaigns going on right now are not measuring properly or are not measuring the right things?
McHale: There are certainly a couple of servers out there who have been late on the uptake to offer more sophisticated tracking capabilities. For us, the question is: Is it a qualitative measurement or a quantitative measurement? If it’s a qualitative measurement, where we’re actually not going to be selling anything specifically, we’re looking at impressions, clicks, click-throughs, CPC, and then any kind of post-click or view-through, looking at the number of users on the homepage, number of users on any secondary level pages. And then after that, is there any kind of desired response? Did they participate in a poll, or did they give us an email or request more information or whatever? I’m not sure you can do from a strict site-measurement standpoint anything other than that. What we’re trying to do is work with the Web Trends, the Blue Martinis, and figure out how the media can be tied in with the back end in such a way that we can really measure continuous response or continuous interaction, repeat visits, and all of those things. So then again, we build up a log and then we’re able to measure how well not only of course the advertising is doing but the site.
Jacobson: I think obviously in some cases they are measuring the right things, but in many cases, from what I’m seeing, they’re not. Of course, it depends on the campaign. If it’s a branding campaign, I think for the most part, you could just throw click-throughs out the window, especially if you’re doing sponsorships and things like that. For a while, media planners have been showing their clients, “Look, we got excellent branding just by showing them a screen shot of a visual sponsorship.” But today, using the different tools out there like Ad Index and some of the other ones, we could actually quantify this branding impact. And, of course, in a hybrid deal, you’re going to want to look at clicks and conversions and some of the behavioral data as well. But I think that people have been looking at the wrong things for a while, and just recently are they starting to change some of the ways that they’re measuring campaigns and look at some other new metrics and measurements. Another one could be people that maybe clicked over to a site and didn’t convert, and maybe looking at what they’re doing and looking to retarget them—exclusively serving a banner to just them, in some of these new ways of targeting based on behavioral information. So there’s a lot of new things out there that I think planners are just starting to adapt to.
Jim Meskauskas, chief Internet strategist, Mediasmith: The bulk of our clients have all been so data driven that the kinds of things that we are tracking are indeed actions. Whether it’s at the view-through level, the impression-based level, or the click-through level, that’s something that we’ve been pretty much doing across the board for all clients, since the client is the one who is going to be the ultimate dictator of the objectives. And it’s great, because traditionally they always said, “Well, you know, I want everyone to have a warm, fuzzy feeling about my brand—and I need 10,000 registrants.” And you have to prioritize the objectives: Is it going to be a direct-response objective or a brand-awareness objective? And then measure based on that.
To date, most of our clients have ultimately settled upon, “Well, we want a direct-response objective satisfied first, and then hopefully there’s some kind of brand effect associated or affiliated with that.” And more and more, the industrywide conversation is finally coming around to where it should have started: The Internet is not like a redheaded stepchild or some sort of character from a V.C. Andrews novel shoved in an attic somewhere. It’s a medium and it should rest on the shelf along side all of the other media. It’s got its strengths and weaknesses like other media do. It should be used as part of a marketing mix, in order to create a kind of Surroundsound marketing environment and maintain multiple points of contact with an existing or prospective consumer.
And only now are clients starting to even think that way. It’s no longer some Balkanized fiefdom in the basement of the marketing department. The Internet is now part of the rest of their media. And the kinds of things that can be tracked are actions—those are still the best things that we can track. More and more they will migrate towards “What’s the brand impact, purchase intent?,” those kinds of things. But some of the questions they want answered first are “What’s lifetime value? How many times has this person returned to my site?” Our publisher clients want to know, “Who comes from what place and generates the most number of pages viewed? Because I’m a publisher and I have to sell ad views and I know how much I can sell them for. How much can I generate them for, and over what period of time? How long is this particular acquired customer going to keep coming to my site and being basically someone who generates page views for me that I can turn around and sell?”
Media: What do you think about the Internet as a marketplace of ideas, where you have people on, say, Epinions.com, exchanging opinions about a different brand and really affecting how one another feel about a specific brand or a specific product? If you happen to hit these people with your Ad Index tool, you could still get a gauge on that effect of how people are responding to the brand online.
McHale: They’re good tools, but when we talk about statistically projectable to a national experience, I think they’re more qualitative than quantitative. I think they’re great test gauges; if you don’t put too much pressure on the results and are looking for some basic metrics, you can get some learning. But when you interview 300 people online and ask six questions…
Media: To be fair, the numbers are better than that, and there is validity to actually setting some benchmarks now to start to track this over the course of every quarter for a brand’s lifetime. I would suggest that any company tay, 92s building a business should be building a brand. And if they are planning on being around for a long time, getting these metrics or reads on brand equity on at least a quarterly basis is necessary.
Meskauskas: The real point is that we’re finally accumulating enough research to start being able to say something about what is or is not possible on the web and how that fits within the context of the overall communications package as far as the rest of one’s marketing efforts are concerned. But I do think that like all research, there’s more directionalism to it than actually answering questions. We all know that sample basis and focus groups—things like the Arch Deluxe and New Coke came out of stuff like that. So we have to be careful when we use research like that. But we are at a point where we can start to make some statements based on something other than licking your finger and sticking it in the air. [Laughter]
Media: If we were having this conversation three years ago, we’d probably be talking about optimization and things of that nature and DR campaigns. But what does this new feedback loop do for the old tried-and-true test/learn/optimize cycle?
McHale: I’m hardly knocking research. We just sold a mega project with one of our largest consumer packaged-goods clients where we did a major deal with one of the portals that are mostly reaching women and we’re testing seven brands over the course of nine months. And we’re actually doing a pre-, a mid-, and a post-analysis—all based on Ad Index, Dynamic Logic methodologies—for each of seven brands. Because this client spends hundreds of millions of dollars in traditional, we’ve stepped up to the plate with a nice deal with one of the portals, and they have their whole research department.
Believe me, the last thing I’m doing is knocking research. I think there’s value to it. But, again, it’s more focus-group-like intent to buy. And quite honestly, what I was very vocal about in this process was being able to focus on getting some quantitative metrics; asking people about participating in a poll, listing their email—any kind of interactive response for me would have been a much more compelling argument, along with the qualitative, to go in at the end of this nine-month campaign and be able to show, yeah, we’ve got all these people talking about moving the needle, but look at all this other interaction. And it only grew when we looked at the log. So intent to buy is great, but any retailer, any company selling its products through mass-merchant supermarkets likes to see overall traffic grow, not decline. So those to my mind are more compelling to brand managers who are very, very cynical or taking a show-me type of view when it comes to taking even 10 to 15 percent of their budget out of television and moving it into the Internet to have a real presence online. At the end of this deal, which will be up around June, we’ll be able to see exactly how it performs. Again, as a sales person for the medium, as an evangelist here, that makes a much more compelling story than only intent to buy.
Oates: I agree with your thinking in terms of interaction, but interaction itself is not necessarily a good thing unless you actually measure what that gets you, whether it’s registrations or whatever. I understand that you want some sort of behavioral metric. But who really cares if people are interacting with something unless you know that that interaction actually made them feel more loyal to the brand or made them want to buy more or made them want to evangelize it? Interaction itself is nothing unless you know exactly what that interaction gets you. Time spent, even—you could have people playing a little game that you created on a Bounty website, and they could be spending hours and hours and hours. But if research came back and said that it was actually eroding the brand or diluting your brand, then you wouldn’t want to do it. And it really doesn’t matter if they were spending a total of 100,000 hours on it if it doesn’t do any good.
Jacobson: One new technique that I could see being used in the near future is not only taking these new branding and attitudinal metrics seriously in post-campaign analysis and things like that, but actually taking these metrics and optimizing on a mid-campaign, as opposed to optimizing on click-throughs or conversions or sign-ups—especially if the objectives of the campaign are entirely branding oriented. Why not optimize the campaign, why not run more creatives that are getting a higher intent to sign up for a newsletter than one that’s getting a lot of click-throughs but not much intent or branding impact?
Meskauskas: The secret, the man behind the curtain of all of these kinds of campaigns is: at the end of the day if you’re not moving widgets, it’s not working. And there’s not a client out there who only wants their brand to be well-thought of. Because if I think it’s really great and I never buy it, the goal has really not been satisfied ultimately, which is: I have to move units off shelves or get people to engage my service. Optimizing against increase in awareness and those kinds of things, if that can then be translated demonstrably into actual widget sales, then I could see doing that. Otherwise, the client at the end of the day is going to say, “How much money did you make for me today?”
Media: What if you’re more of a traditional advertiser that’s not selling product online, but your advertising online brand manager would basically live and die by a lift or a decrease in purchase intent?
Meskauskas: But look at Taco Bell. The reason they were going to fire FCB again was, yes, their ads were really cool and everybody knew who the chihuahua was, but they didn’t sell more burritos. Even television—let’s take all that Scandata and run it against schedules to see whether there was a lift in sales.
Media: I think you just hit the nail on the head. That’s where I think we’re going. When a research firm can place cookies on a certain group of their household panel, 7,000 or 10,000 people from their household panel of 55,000, and watch their offline purchases and try to see how digital marketing campaigns affected their purchases. And there’s some more information that’s been coming out about these tests, and they really do show at least for certain types of products that online marketing really can boost overall sales.
Meskauskas: And I think that’s the kind of thing that will need to be demonstrated in order to see the web as an ad medium grow and become a true, honest, and accepted part of the media mix in advertising. That sort of information—being able to show that correlation—is going to take the web to the next level where it needs to be. Oates: We have a huge advantage and we’ll continue to exploit it. And from our end, the sooner we can actually see how we’re moving offline sales and online sales together—track cookies across multiple different sales channels or across multiple different marketing solutions like search-engine optimization, or precision email, or consumer targeting or strategic partnerships and generally standard web media—and look at how all of our digital marketing mix contributes to moving the bottom line, the better it will be. Nowhere else except in digital media do we have that opportunity.
