By David L. Smith, President/Media Director; Mediasmith, Inc.  smith@mediasmithinc.com

A version of this article was first published by MSN Advantage in Q1 of 2003

I have written about Reach and Frequency a lot over the past few years. Lately, there has been some discussion at various conferences (@d:tech, iMedia and others) as to whether this topic was a) done b) worn out c) had a long way to go or d) not even relevant. The last is the easiest to deal with. The critics say that, to do R/F for the Web is to put ourselves at the mercy of the TV planning model. I don't think that's true. R/F is an important metric for predicting impact of campaigns and tracking their efficacy. Just because we have R/F for the Web does not mean that we need to be automatons and only use it in computer optimizers. Sometimes clients just want to know what the Web adds to their other campaigns in impact, and they have a right to know.

As to the other comments, it is not done. Far from it (see below). If you think it is worn out, then you are more concerned with style and this year's model than appropriate media metrics. The correct answer, unfortunately is "c", we have a long way to go.

In past articles in this space, I have talked about the work that is being done by a number of companies to produce reach and frequency systems. Most of the work that is being done today, especially the already in-market systems from IMS using NetRatings data (WebRF), the DoubleClick MediaVisor system using the same data, the Atlas DMT GRP and RF Tool, the comScore tool, the Telmar Webplanner system, BlueStreak and others concentrate on evaluation of alternatives at the planning or pre-buy stage. But, as I have stated before: "In the end, all of these programs are only one part of the needed pie, the planning and pre-buy part. Final releases from the key vendors will need to have the ability to track the reach and frequency of a campaign as it is in progress as well as a post-analysis. After all, in the end, the proof of whether a reach and frequency estimator is accurate and provides the correct algorithms and data inputs will be established through post analysis. If the actual post buy R/F is equal or relatively close to the pre-buy R/F estimate, media planners, buyers and sellers will have what they are looking for."

In fact there are four specific applications of reach and frequency for the Web that are currently being talked about by senior agency media managers. These include:

• R/F at the strategic level
• R/F at the planning or pre-buy level
• Tracking R/F delivery as the campaign is running
• R/F on a post-analysis basis

I'm going to spend some time in this article discussing how these four areas are related and how they could play out in future software offerings.

1. R/F at the strategic level: This application would permit quick, broad stroke strategies to be developed much the same way we look at other media strategically. The reach and frequency data for these strategies would then be combined with reach and frequency data for other media either manually or through the use of an optimizer. The first two companies providing offerings are IMS and Telmar. IMS and Telmar have historically been used by agency media operations. They make resources like MRI for print available for cross tabs, efficiency analyses and R/F comparisons within traditional media and on a cross media basis. Their TV R/F capabilities are based on the Nielsen TV cume study and their radio capabilities are based on ARB. IMS has a greater customer base with U.S. agencies but Telmar has a healthy U.S. market and is stronger overseas. To compete with the already in use IMS WebRF, Telmar has developed WebPlan for Web cross-tab analyses and R/F. Both services have very mature cross-media analysis capabilities that permit media planners to compare media and combine them in R/F analyses. Media planners will use these capabilities heavily in order to show clients how much Web activity should be used in the media mix, and where the strengths of the Web are demographically to complement the current media mix. Whether Web should be planned this way or based on its merits alone is a subject for another day. But we believe that this tool, combined with the maturation of rich media and the availability of Web and cross media awareness studies from companies like Dynamic Logic represent the future for establishing the efficacy of the Web in a branding media plan.
   
   
2. R/F at the planning or pre-buy level: The offerings in the market today from IMS, Telmar, Atlas and comScore bring this capability to the desktop already. Some day "real soon now" the media agency and the seller will be able to choose their interface and their data bases to use in examining alternative plans and buys. (See my R/F articles on MSN Advantage to understand more about these offerings.)
   
   
3. Tracking R/F delivery as the campaign is running: We know that the Web is unique in that we can measure many things not available for other media. Tracking reach and frequency delivery as the campaign is running is not currently available but the technology clearly exists or can be tweaked to do this. The third party ad servers already have such tools in place from an overall impression standpoint. Services like Atlas DMT and DoubleClick's MediaVisor, working with UCMs (user centric measurement companies like comScore and NetRatings) should, in the near future, be able to provide buyers with the ability to watch the R/F of a buy accumulate. It will then be possible to understand whether or not the sites are allocating inventory the way the buyer anticipated they would. The buyer would have the ability to go back to the site and instruct the site to provide a broader rotation (to raise the reach accumulation) or tighten up the rotation (to raise frequency levels) while the buy is going on. This has never been possible for any other medium. In this manner, R/F optimization is born as the sister of back-end optimization based on CPA for DR campaigns. I recognize that this service is not trivial. As of this writing, Atlas DMT does have a tool available for tracking online campaign R/F accumulation but it is such a data crunch that it is not available in real time, but 9 days later at the earliest. But then I remember in the 80's when I had to take a break and go to lunch while my computer was crunching something that I can now do immediately. So hopefully we will resolve this problem in time and get these products in near real time, possibly with overnight processing of batch files.

   Tracking R/F accumulation while the campaign is running is especially important as the much-ballyhooed concept of frequency capping for a campaign has been uncovered to be undeliverable. Many of us have bought into the concept of frequency capping based on sales pitches by companies like DoubleClick. But the reality is that their technology and that of others is based on frequency of exposure of a single ad, not a campaign. This is better control than we have on other media but is not what it seems on the surface. Once that ad reaches it's cap, another ad from the same advertiser is substituted. Thus, the overall campaign can have no cap under the current system. Individual sites also can cap frequency for a specific advertiser's campaign if they serve all of the ads. But in order for frequency capping to be executed across a whole campaign of multiple sites, a universal cookie will need to be deployed. This concept could be executed but it would need to have cooperation from all of the major vendors and carefully constructed so as to avoid any privacy issues.
   

4. R/F on a post-analysis basis: This capability is already in use by most third party ad servers (3PAS) on a gross impression basis. Overlay of the UCM data and parsing the data for individual country geography is all that remains to put this data on the same basis as the planning/pre-buy analyses. (I know, easier said than done.) As has been stated before, herein lies the proof in the pudding. The debate as to which data bases are precise (within the acceptable range variations of media research), whether SCM and UCM data must be combined to get reliable predictive R/F data and which vendors have the right overall algorithms in place will be born out by post-analyses. If the actual campaign results are consistent with the planned/pre-buy R/F, then we will finally have a system in place that can be counted on to serve us in the years to come.

There remain of course many points of debate among those in the industry about the movement to develop R/F standards. As mentioned above, some Web centric media people claim that using R/F lowers the Web to the standards of traditional media and does not characterize the Web potential in its best light, as a vehicle that can accomplish branding and response at the same time. While the above is true, I believe that the push towards a complete R/F system as outlined above will only benefit the Web. It will bring in more advertisers. It will make it easier for the traditional media planner to consider the Web in the overall media plan, rather than have the client allocate a separate budget to be executed by a separate organization. It will demonstrate to all where Web media strengths lie in respect to other media. It will permit both sellers and buyers to efficiently look at alternatives that best accomplish the stated client goals, whether they be reach, frequency or providing both against a demographic or other defined group which is under-delivered by the current traditional media mix (much the same way that cable TV took off when the data permitted analyses that broadcast TV underdelivered in major markets where cable could provide delivery, thus resulting in more even delivery on a national basis as well as higher R/F delivery overall because of superior efficiencies).

What's next? I suggest that you get the tools on your desktop and test them. All of the vendors (see my articles on MSN Advantage for a more complete list) want to sell you their system. But most will permit a short desktop trial against real business so that you can see the benefits. The market needs to speak out and determine which of the systems work best for their needs. While we don't yet have all of the capabilities outlined above, we do have robust strategic, planning and pre-buy tools that merit attention. Only through active participation in the process can you ensure that you will get the right tools for the future. To that end, Mediasmith is going to test as many of these options as are made available and hope to talk about each of them in greater length in a future article. Watch this space for more news.

David L. Smith is President of Mediasmith, Inc., the Integrated Solutions Media Agency based in San Francisco. He can be reached at smith@mediasmithinc.com.

You are receiving this newsletter from Mediasmith as part of our efforts to keep our friends abreast of what's going on in media.

If you'd rather not receive this or any other Mediasmith publication in the future, you can unsubscribe by e-mailing us at newsletter@mediasmithinc.com with the word Unsubscribe in the subject.

If this issue was forwarded to you and you would like to begin receiving a copy of your own, please visit our site - 
http://www.mediasmithinc.com/contacts/subscribe.html to sign up for these newsletters.