Media Metrics: ROI, the New Paradigm

My response to John Wanamaker's quip about not knowing which half of advertising is wasted: I don't care which half is wasted; I want to know what my sales per advertising dollar are. There's a lot of buzz these days about return on investment (ROI) on marketing expenditures. ROI is the new evolution - or revolution - in the offline world. But one group, namely direct marketers, was squarely centered on ROI long before the Internet began.

ROI means different things to different people. I don't know why a simple concept like return on investment should require apocryphal interpretation. ROI shouldn't be a new and difficult science. Back in the '90s, it seemed that the solution involved the perfect mix of reach and frequency. Media plans came in demanding that at least half of the reach goal should be at the 3+ level. (Oddly, it seemed that all plans, even plans for very different products, required the same mix.) If that balance was achieved, the heavens would align perfectly and move product off the shelves or cars off dealer lots. Agencies would know this because the ratings would post precisely, and a good balance would be achieved.

Where did the 3+ frequency idea come from? Were the first two reaches good for nothing? Was the fourth just overreaching? The idea of the "Magic of the 3" was based on a brief empirical study published by educational psychologist Hermann Ebinghaus in '85. That's 1885, mind you - well before cable advertising took off.

Ebbinghaus' conclusion: If you expose experimental humans to lists of gibberish and measure their recall, recall improves most dramatically after the 3rd exposure. Eureka! Rowe's conclusion: Media investment should not be based on 19th century gibberish.

Fortunately, we are working to find the perfect marriage between brand awareness and direct response. But ROI shouldn't be subjective. Return on investment means, "How many incremental dollars did we receive based on our investment in media?" It should be straightforward. The agency and the advertiser ought to have a simple and clear partnership over this formula. It works like this: The advertiser knows exactly how much revenue it generated in similar periods before it hired the agency. If the agency can tell the advertiser how much revenue was generated in response to agency advertising, and the advertiser can subtract base level revenue, then ROI is self-evident.

The simple formula is: (Incremental $ In) / (Advertising $ Out) = ROI. This doesn't feel much like 3+ frequency, in fact it actually supports recency theory. We know this method works because we can measure it. At Carat Interactive, we use a combination of syndicated and proprietary tools to deliver metrics. These include the syndicated CoreDirect, Marketing Management Analytics, and proprietary Carat processes and analytics.

CoreDirect was developed by COREMedia Systems Inc., which is a direct response management and analysis company. It tracks where our clients' spots run with orders via phone calls or Web site hits. The product supports as many different pieces of lead information as we want to analyze, tracks the entire buy-side workflow, and offers sophisticated tools for reporting and presentation, including graphing and mapping. We worked with CoreDirect to add proprietary logic for sourcing vanity numbers (like 1-800-jenny20) that apply to Web site data as well.

The thing I love about interactive commerce and direct response TV is that it's so concrete. You can see the clicks and incoming calls in real time. With the right tools, you can tie this data to Web spending or tv spot clearance. You can watch the cash register ring and tie it back to the media spending.

The thing that I love about vanity numbers is that the name and the product become one; it's the best parts of direct response mixed with old-fashioned branding. But vanity numbers also allow fuzzy math to creep in. If you run a unique 800-number on every station on which you are advertising, you can easily tell where the calls come from. But unique 800 numbers are hard to remember and they diminish the brand. Vanity numbers like 1-800-jenny20 are easy to remember. But if every station is running the same number, how do you know where each call came from?

To solve this problem, we worked with the CoreDirect programmers to compile a call database by specific creative execution, one that segregates call bursts and defines and applies call attribution envelopes. This allows us to build data on how specific creative executions perform by station, and the data generates a unique call attribution algorithm for each piece of creative. The algorithm adjusts dynamically with changes in the schedule and call burst rate.

We also developed a module with CoreDirect that allows us to track the value of a customer over the sales cycle or the customer lifetime. Once the initial customer contact is established, via call or click, we tag that customer with a unique ID. Then, CoreDirect tracks every contact with that customer and ties it all back to the initial media spend that lured the customer in the first place.

For example, take the case of marketer Jenny Craig, where there is a lifetime cycle of graduated purchases. Tying downstream purchase activity back to the initial media spending has proven invaluable.

It's not that measuring the value of a customer over a period of time is new. It isn't. But by integrating customer identification with lead data that we obtain via the course of our normal direct response workflow, the measurement is much more timely and precise. CoreDirect technology enables each respondent to a TV ad his/her own unique identifying tag. Those customer identification tags track individuals each time they contact an advertiser, even if years lapse between contacts. Having better data enables us to make better decisions. In direct response, we don't post-analyze the campaign to see if we hit a metaphysical goal. We manage the campaign in real time. The goal, of course, is positive ROI.

I can't take full credit for these helpful tools. I work with analysis and modeling tools from Carat's Marketing Management Analytics unit to improve our understanding of direct response media. MMA's modeling helps us to compile the data to analyze base-level sales.

For Jenny Craig, a client of Carat for more than eight years, we are in the process of building the brand while simultaneously providing call volume. In fact, Jenny was one of the first advertisers to approach its media from a blended perspective.

We read results weekly and make adjustments to our schedules in order to optimize and offer the largest amount of leads. MMA's data helped us to refine our media mix and concentrate on media that provided the greatest ROI for Jenny Craig.

We are always looking for ways to mine the data and go beyond what clients expect. If it means teaming up with technology companies to scope out the next wave of innovations, we do it. Advertisers are demanding that their spending show a solid ROI, and that means concrete results instead of abstraction.

Susan Rowe is executive vice president, integrated media, at Carat Interactive. (

Next story loading loading..