In This Issue:

Volume I March 2008

How Targetbase Turns Data into Profits

Telecommuters Profiled

When is it Better to Report Data in Terms of Median or Mean?

Your brand sells off the shelves in Bakersfield, but collects dust in Bangor. Your ads are looked at in Providence but overlooked in Portland. We've all known for a long time that what works in some markets -- or nationally -- does not necessarily work in other markets. Now MRI's Market-by-Market study provides information to tell you why and intelligence so you'll know what to do about it. 

Market-by-Market delivers at the local level the same robust and reliable data that marketers have come to expect from MRI on the national level. Just like MRI's national Survey of the American Consumer, Market-by-Market delivers information about consumer behavior for approximately 500 product categories, encompassing 6,000 specific product or service brands. If you're looking for information on demographics, consumer actions, media usage, exposure, psychographics or volumetrics, Market-by-Market provides it for each of the 205 DMA's (Designated Market Areas) in the continental U.S. Even consumer segmentations drawn from attitudinal and behavioral questions in the survey are included. 

Market-by-Market also enables custom regionalization; marketers can build up -- or break down -- custom marketing regions and no longer are limited to pre-defined census regions, marketing regions or states. 

Using MEMRI, MRI's proprietary data analysis program, anything that can be done with MRI's national survey data can be done at the Market-by-Market level. The full range of analytical and reporting tools is available: cross tabs, Golddigger, mapping, autocoding, N-tile and frequency distributions, trending, charts, slide shows and presentations. MRI Market-by-Market data will also be available through IMS, New Age and Telmar systems.

Market-by-Market - Magazine Types at a Glance 

Sample of Types of Magazines More likely to be Read in Select Local Markets

Index = 100 - Total U.S. Adults

Source: MRI, 2007 Market-by-Market

Learn more about MRI’s Market-by-Market study. 

back to top

This Commentary on issue-specific audience data was reprinted from the launch issue (1/27/08) of CircMatters, a newsletter published by Jack Hanrahan. CircMatters is about the business side of magazines. It's for salespeople and circulators, for agency strategists, researchers and buyers, for advertiser clients…and for anyone else that touches the ever-changing world of consumer magazines. Previously, Mr. Hanrahan has held key decision-making positions on both the vendor and client sides of the business, having worked at Leo Burnett, Coca-Cola, Prime Point Media and OMD.

You Asked for It…You Got It…Where are Ya?

CircMatters figured that, if the title said "Issue-by-Issue Audience Data", you'd just pass it by. After all, that's what most agencies are doing with the new issue-by-issue product from Mediamark Research & Intelligence (MRI). That's right…the same more granular, faster information they have been asking for. Let's back up at bit.

In forum after forum, many agencies have pushed for data that would make magazines more comparable to and competitive with other media. A big part of the "ask" was getting audience data on every issue published.

This would take magazines to a place well beyond an average issue audience. In April of 2006, in response to the marketplace (publishers, advertisers, and agencies), MRI announced a self- funded pilot study of issue specific audience measurement. Self-funded….as in "MRI paid for it themselves".

That pilot evolved into MRI's commitment to a continuing web-based Issue Specific Study. In June 2007, MRI hit the streets to sign up subscribers for this new product. For the most part, support has come from key publishers. To date, only two agencies are on board. Kudos to Starcom and MediaVest for putting their money where their mouths were.

Where's everyone else?

It's time to get on board.

Sure, the data provided are indices and those indices (right now) are solely on adults, men, and women.

Too limited? No way…not at this early stage of what could and should be a turning point for magazine measurement. There's a lot to learn from what is out there right now.

Need convincing that the data released by MRI to date enlighten?

Look no further than a paper delivered at last fall's Worldwide Readership Symposium in Vienna by 5 U.S. researchers from Time Inc., Mediacom, and MRI (see note for authors and a link to the paper, below).

When the study was done, MRI had released about a year's worth of data. Here are some of the interesting and important questions the paper explored:

• Do high reader-per-copy (RPC) titles (like Brides or Soap Opera Weekly) have more variability between their highest audience and lowest audience issues than do low RPC books (like Working Mother and Family Fun)? 

Finding: The two groups had about the same "index gap" from high to low performing issues.

• Does the cover subject of People magazine influence the total audience level? 

Finding: It seems to. Covers dealing with tragedies like the Virginia Tech shootings, the Amish School Massacre, or kidnappings tend to index below average. By the way, this below average audience delivery for "tragedy covers" wasn't evident for newsweeklies; their Virginia Tech shooting issues performed on par with the average issue for Time, Newsweek and U.S. News.

• Does the male/female skew of a title shift with the cover subject? 

Finding: For Entertainment Weekly, the paper reported "movie covers did well among men while covers showcasing TV celebrities or TV shows performed well among women." And everyone seemed to like Johnny Depp (or pirates) as you can see from the covers below.

There's lot more in the paper. Check it out. Re-think where you stand on this issue-by-issue initiative. It's time for those who haven't subscribed yet, particularly the agencies, to jump in the pool and start to play. Data is coming on stream allowing giving those with access the first chance to see what seasonal variations there are (or not) by title and genre.

Without more agency support, will this new resource continue?

Without the data, will you be as smart as your competition in the next big new business pitch for a big magazine spender prospect?

Do you want to take the chance on the answer to either of those questions to be "No"?

Read the paper written by Caryn Klein & Lori Jacobs, Time Inc. Alan Rovitzky, Mediacom, and Mickey Galin, Julian Baim, and Marty Frankel, MRI.

If you'd like to read the first three issues of Jack Hanrahan's CircMatters, or if you wish to subscribe, go to CircMatters.com.

back to top

Successful direct marketing strikes a chord with consumers and yields high response rates and satisfied customers.  But predicting what will grab consumers’ attention in an over-saturated media marketplace is a growing challenge.

Targetbase, a direct marketing agency headquartered in Dallas, is addressing this issue by using modeling analytics to create a database that provides a comprehensive array of actionable direct marketing intelligence.  In this Q&A, Targetbase Vice President Product Development and Innovation, Bill Cole, discusses the development of this database, done on behalf of a consumer goods client.

The nature of this client’s business does not involve a transactional link with its customers. Without the customer raising their hand in some way, or responding to a direct solicitation or offer, this client has no idea exactly who is and is not using their products.

The client did, however, possess a fairly large database of consumers and prospects, which was sourced from a diverse set of product groups and brands as well as from multiple 3rd party data suppliers.  But, the “intelligence” value of the data was marginal and fragmented which limited the database’s value for strategic marketing. 

Our client needed universal and strategically actionable data to both broaden and enrich their database’s value for direct marketing.  They hired Targetbase to develop and implement a solution using predictive analytics as a means of creating these data. 

Q. Please tell us about how this solution was developed and what information it provides?

This project really called for a custom solution from MRI. The types of actionable data needed were very broad. MRI’s syndicated survey data offered the best and most economic means of accomplishing the predictive analytics required because it covers a broad array of consumer characteristics, behaviors and attitudes. 

MRI’s Market Solution group worked out a plan for the creation of a special modeling file, which featured MRI respondent-level data matched to corresponding records on our client’s database.  Everything was blinded to ensure the privacy and security of each party’s data.

The end result was an analytic file that was ideal for developing the array of statistical models we needed -- a large sample of anonymous consumer records for which we had both the information available on our client’s database and their responses to the MRI survey.

It’s not unusual to combine proprietary client database records with survey responses for analysis and modeling purposes.  But, to have database records combined with as large and diverse a survey as MRI’s definitely is unusual.

Within this combined dataset, we can now identify consumers who, for example, are heavy users of a certain type of product.  We can then develop a statistical model (or algorithm) to determine any individual’s likelihood of being a heavy user of that type of product using only the data elements contained on our client’s database, such as demographic information, psychographics and response history. Once this algorithm is created, it can be run across all consumer records on the client database to create new selectable data on each consumer’s record -- specifically, a probability score reflecting their likelihood of being a heavy user of the product.

This is the basic premise of the project -- the development of new data, representing predictions of product usage, brand preference, retail behavior, media receptivity, offer receptivity, etc.  The process is simply repeated over and over again. To date, we have created more than 100 of these models, and their associated predictive elements.

So, instead of being merely a list of consumers, with widely scattered knowledge and insight on each one, the database is now becoming a rich resource with an increasingly comprehensive array of actionable direct marketing intelligence.  

Q. How is the database being integrated into your client’s direct marketing plans?

The database is a tremendous asset for the client because the intelligence it contains is robust and highly actionable for direct marketing.  It’s more effective because the type of intelligence we’ve created has been carefully selected to drive better planning and targeting.

It also saves them money.  Product divisions are no longer forced to seek custom data suppliers for targeted direct marketing.  It’s all consolidated now, so it’s much more economic, not to mention more convenient.

Q. Did this project result in the building of protocols or processes that you’ve extended to other clients?

We are assessing opportunities to extend this type of solution to other clients.  The integration of direct and mass marketing intelligence is increasingly important to our clients, and the integration of MRI data and client databases, combined with the power of predictive analytics, provides that kind of integrated perspective and insight.

Targetbase is a direct marketing agency that specializes in providing a broad range of clients  actionable insights that turn customer data into profits. Founded in 1979,Targetbase is part of the Omnicom Group. 

Learn More about MRI's Market Solutions Group.

back to top

Working from home can be ugly. Cluttered desks, half-empty coffee cups all over the place, single socks scattered on the floor … people who don’t shower until noon -- if ever. This is life for many of the 4.3 million adults who telecommute.

According to MRI’s Fall 2007 Survey of the American Consumer, nearly 2% of U.S. adults work primarily from home as company employees. Thanks to technological advances such as instant messaging, text messaging, emailing and video chatting, they can easily join co-workers in meeting complex business goals across a range of industries.

Here are a handful of telecommuter insights revealed in MRI data:

• More men than women telecommute -- More than one-half (52%) of telecommuters are male and 48% are female. 

• Telecommuters have enough on-the-job experience to work without the structure of a traditional office setting -- Four in ten telecommuters are between the ages of 35 and 49.

• Most telecommuters work in sales/office occupations -- More than one-third (34%) of work-from-home employees have sales/office occupations, 21% work in management, business and financial operations and 19% hold professional and “related” positions.

• Telecommuters are more likely than the average adult to be well educated -- Compared to the general population, they have an index of 175 for having earned a bachelor’s degree and 161 for having earned a post-graduate degree.

• Work-from-home employees live in high-income households -- Telecommuters index 209 for living in households with HHIs greater than $150,000.

“Telecommuters are an attractive target for a wide range of products and services,” says MRI Vice President Marketing and Strategic Planning Anne Marie Kelly. “From business services geared towards people at home -- such as independent IT support -- to local leisure services, they have key demographic attributes that make them a segment worthy of targeting.”

Got it Made! Telecommuters have the flexibility to exercise, 
eat out and entertain, plus some:

Compared to colleagues who work at the office, telecommuters are more likely to: 

Index = 100 - Total U.S. Adults

Source MRI Survey of the American Consumer Fall 2007

back to top

When comparing demographic variables, for example household income or age, is it better to report the data in terms of median or mean?

Let's begin by reviewing the difference between median and mean. In a median measurement, the data are arranged from lowest to highest with the middle number in the set being the median. For example, in the set 10, 20, 75, 80, 90, the median is 75. When considering mean, however, the numbers are added and then divided by the numbers in the set to get an average. In this example the mean is 55.

The mean is often affected by outliers; where the data are farther removed from the central tendency of the number. For example, if someone is looking at the Individual Earned Income for all adults, a small number of especially highly paid people can skew the mean higher. For example, the mean IEI for U.S. adults is $44,722 while the median is $35,414. If you are looking for an overall average number to represent the population, use the mean. But if you want all the IEIs in the demographic to contribute equally to the analysis, go with median. 

Which is the right measure for you? It depends on what information you want to emphasize. If, for example, you are developing a presentation demonstrating how your publication delivers more upscale readers than your competition, running mean data will produce an average where the higher income earners will tip the scales to reflect their presence in your audience. The median will provide a snapshot of the income of your audience without tipping the scales in either direction.

back to top

min Guest Commentary by Dr. Julian Baim

MRI EVP/Chief Research Officer, Dr. Julian Baim, authored a Guest Commentary for the March 10th issue of Media Industry Newsletter (min) discussing learning from MRI's Issue Specific Readership Study. 

Dueling Covers: A Look at Issue Specific Readership

Pew Trust's State of the News Media 2008 Report Highlights MRI's RFID Progress

MRI President & CEO, Kathi Love, as well as other industry leaders, shared their thoughts on the state of media, advertising and measurement for a section of the Pew Trust's State of the News Media 2008 report. Read here.

back to top

The Source
Mediamark Research & Intelligence
A member of the GfK Group
75 Ninth Ave 5th Floor, NY, NY 10011
www.mediamark.com           www.mriplus.com

New York (212) 884-9200 • (800) 310-3305 • Chicago (312) 329-0901 • Los Angeles (323) 882-6325

Comments? Want to be removed from The Source mailing list? E-mail and let us know.

^©2008 MRI All Rights Reserved