• by May 22, 2002

The Media Debate
Let’s say your media target is men 18–34 and you’re about to buy a million impressions on site X to reach them. Immediately, you’re faced with two issues. The first is that many media minds believe that if any of those impressions are seen by teenage girls, they are wasted expenditure. The second issue is that there’s a big difference in advertising impact between delivering 500,000 impressions once and 100,000 impressions five times. So how do you know how many different people you’re really going to reach, and with what frequency to reach them?

It’s "Reach and Frequency Meets Online Audience Measurement," and it’s one of the most important topics of discussion among online advertisers today. So important, in fact, that the Advertising Research Foundation recently put together an Online Reach and Frequency Committee to come up with a plan of action. The committee met in relative secrecy last month to decide which audience measurement solution is the best for establishing online reach and frequency guidelines, and MEDIA has invited some of the participants of that meeting to shed some light on the topic.

Moderator - Tim McHale, former Chief Media Officer, Tribal DDB: There has been a limited amount of information from the ARF meeting, some positive and some negative. Could we go around the room and share some insights as to what your take-away was and what the ramifications are?

Steve Coffey, Coffey & Associates: I think that meeting really marked the beginning of a change in the way people are thinking about the role of audience measurement data in support of advertising on the Internet. It used to be that audience measurement data were used primarily for promotional purposes — to go into brochures and to take your company public, back when that was fashionable. Now people are trying to figure out how to make the ads perform, how to make the ad buying process efficient, and how to use the available information sources in an organized way. I got excited about seeing how individual and whole-industry efforts are starting to move forward, starting to plan online advertising like it’s really advertising rather than promotional hype.

Doug Knopper, Vice President and General Manager, Advertising Solutions, DoubleClick, Inc.: I felt like people walked into the room [feeling] very proprietary, feeling that their solution was the best. And I think we walked out of the room saying that there is a bigger industry solution here that we can all provide. And it felt like there was a collective understanding that the importance of reach and frequency is bigger than any one of us, and it really has the opportunity to make our industry completely successful.

Jeffrey Graham, Vice President, Client Development, Dynamic Logic: History will tell, but it felt like a watershed moment in that there was a spirit of cooperation, and the online advertising industry has certainly witnessed its share of dissipating battles and arguments and infighting over measurement standards. The thing that came out was everybody’s willingness to find something that was useful for everyone as opposed to just battling over whose metric or whose measurement was the most valid.

Coffey: On the other hand, in that meeting my blood pressure blew the roof when the suggestion was made that we were going to create this socialist solution, that there is only one best way to do this and we’re all going to jump on that bandwagon. I think it’s way too early for that, and being a born-and-bred capitalist, I think that the best thing to keep this moving forward and healthy is that there is some competition, so that different parties are always trying to race the next best one.

Knopper: Steve, I completely agree with you on that, and don’t misinterpret my comments. Someone at the end of the conference said, "Why doesn’t everybody put their products out and let the marketplace decide?" and I think that’s what’s going to happen and ultimately what has to happen. But what we’re all thinking about is what the overall product requirement should be, and on a much broader scale than just "it’s going to be panel-based" or "it’s going to be server-based." The downside of that is, if we all walk out of that in lockstep and come up with one product, then we probably have done a disservice. Young-Bean Song, Director of Analytics & Atlas Institute, atlas DMT: I think everyone’s realized from the discussions around standardizing impressions and clicks and all this kind of stuff is that we’re going to run into the same issues with reach and frequency. I saw the ARF meeting as being an opportunity for everyone to see what everyone else is doing and actually have the conversation that we all wish we had four years ago about what an impression is.

McHale: The entire advertising industry feels that the weight of media allocation is so unbalanced and that we don’t have the tools yet to properly understand what the correct mix is. And because dollars are so scarce and breaking through is so important, reach and frequency is being taken much more seriously, so that we can hopefully counsel our clients to get to the point where we feel much better about the specific media mix, if only from a quantifiable standpoint.

Graham: I also think there’s a tendency for people to look at reach and frequency as a panacea for online advertising’s woes. It’s certainly going to be useful to have a currency that’s comparable across different media platforms, but the main problem in the online industry isn’t that we don’t have reach and frequency models, but rather that all marketers don’t believe that online advertising is going to solve their marketing problems.

McHale: Let’s talk about the four pillars of what goes into reach and frequency. From my read we’re looking at part one, the media audience. Part two is the duplication of media, or how it’s measured. Part three is the media buyer, the media schedule itself. Part four is the target definition — how the marketer describes and defines the audience they’re trying to reach. Are there any other elements that need to be added to this list?

Song: I think the only other one might be time, some notion of serving a million impressions in a two-week time span versus a two-month time span, and how you may garner a different reach/frequency cure.

Coffey: I think that’s been a vexing problem for the Internet going way back; we don’t really know what we’re measuring. In thinking about the reach and frequency problem, I started off thinking about television. When you place one ad on a TV show and you use the average minute audience rating of one million as your estimate, you get a million viewers, not 500,000 viewers two times. In television, the placement of a single ad has no duplication. On the Internet, if you buy a million impressions, you really don’t know whether you’re getting a million people or 500,000 people. And a lot of that can be driven by the duration over which that single ad placement is being preformed, but the complication starts at that very first stage. In media measurement it was always, "You have to measure the vehicle and the ad follows." Well, the vehicle for the Internet turns out to be a page impression by a person. So the ad buys really do start at an incredibly granular level, and I think we missed the significance of that in trying to think about how we support advertisers.

McHale: It sounds like we’re getting into the multimedia issue here. How are these metrics, regardless of how they are calculated, going to be used in the field with other media measurements, other media reach and frequency?

Song: When I hear this question about how marketers are going to evaluate reach and frequency across different media, it’s important to note that offline marketers are already doing that with print, outdoor, and TV. The major Consumer Packaged Goods companies have their own proprietary models which evaluate a GRP of TV versus a GRP from print. Thus, they use discount factors like “one GRP on TV is worth .4 GRPs of print and .2 GRPs of radio.” Whatever it may be, they are able to use those factors to get an overall view of GRPs (and cost per GRP), which allows them to make the proper tradeoffs. Those same factors will be developed for the online medium, and even associate different values to different ad sizes.

McHale: How does the group feel about the mixing and matching of different databases? Coffey: Well, wait a minute. Before we even go there we have to understand why is online not in those models yet. It’s because we don’t have things like reach and frequency and GRPs and TRPs as part of the vocabulary for online advertising. That’s why it’s so crucial that people who are developing solutions get cracking and we get some things out there, because then we can just roll over and be merely another medium on the list for the planning tools that are out there in intermedia.

Knopper: The people that are in this room and the other companies that are working on this are not going to provide all the answers. Much of it is incumbent on different agencies’ providing different models. Just as there’s probably no accepted standard right now for how you measure a 30-second GRP versus a 15. Every agency does it slightly differently and provides its own weighting factors. And some of that is going to be the Tim McHales of the world saying, "Here’s how our agency is going to do it." And a lot of this is going to be a process where we have to provide the base methodology and then everybody’s going to interpret it to whatever meets their needs.

McHale: Steve, you said the industry has to start looking at reach and frequency GRPs, TRPs. Do we believe that the online industry a year from now, universally, or well over 70% of the industry, is going to be quantifying their schedules based on these metrics? Are we talking to ourselves or do we think the industry’s going to accept this?

Coffey: First of all, let’s talk about the Internet and agree that it is a Balkanized advertising solution. When it comes to display advertising on content sites, reach and frequency is the only way to evaluate and figure out how it’s working. If you’re trying to create changes in perception and attitudes, click-through is meaningless. It’s about getting the impression in front of a person and then measuring the value of that impression to see if it works. And to do that today is like selling small publications that aren’t measured — it’s just a content sale and a couple of martinis over lunch to close the deal. You’re not dealing with any reality, it’s whatever you can get. We need to do better than that and deliver answers to the questions such as "How many am I going to reach of my target audience to place a value on it?"

Song: Don’t forget the publishers. The publishers are probably as excited about this as anyone. Because all of a sudden there’s a way to monetize and get paid for the 99.7% of impressions that never get clicked on. The advertisers are excited about reach and frequency, and the publishers are going to be pushing this along as much as anybody.

Graham: I think you can look to traditional media as a guide here because MRI provides magazine ratings data, and that’s their business. Their business isn’t the reach and frequency model, although they have one. If we separate the business interests of the developer/owner/business person who has a vested interest in the model from those of the person who’s got a business interest in the data, we’re probably going to wind up with a better solution rather than having the data owner trying to push that data through the model even if it’s not optimal. There really are different data types that need to be involved to create to best possible answer to the question of reach and frequency.

McHale: I think the dual approach of user-centric and site-centric data opens up the whole idea that if we are going to use two different measurements platforms, this means that all of the traditional media is only half measured. So with that, thank you so much to everybody.