Dotdash Meredith® has provided effective statistical modeling services to

qualified clients for over 20 years. Their experienced staff of analysts

employ proven statistical tools and techniques to create the lift needed to

make marginal lists perform more profitably.

All their statistical services are initiated by matching 100,000 of your best

direct mail responsive customers to the Dotdash Meredith Database.

This file includes the subscribers of:

All Recipes®, American Patchwork & Quilting®, Better Homes and

Gardens®, Coastal Living®, Cooking Light®, Country Home®, Do It

Yourself®, Food & Wine®, Midwest Living®, People®, People Royals®,

Real Simple®, Southern Living®, Traditional Home®, Travel + Leisure®,

and Wood® (+ Meredith Book Buyers)

TYPES of MODELS

1- Good Customer Match (Profile) Regression Model

Send a file of 100M best customers to Meredith.

Meredith will match your best customers to its file and

first use decision tree analysis to develop a customized

universe from the Meredith Database that is more

targeted for your product. They then use a regression

analysis on this universe to determine which variables are

the best predictors for identifying Meredith subscribers

that "look like" your best customers. A scoring equation

is developed which is used to select names for roll out. A

detailed model report, including model variables, gains

chart, and ordering recommendations, will be provided

upon completion of the build. Turnaround time for

model build: 4-6 weeks

Dotdash Meredith Database - Custom

Modeling

2- Mail/Match Model

Send the mail files from some recent campaigns with

responses flagged and, if appropriate, donation or order

amount. Meredith recommends sending mail files with

at least 20,000 responders in order to have enough for

modeling. These files are matched to the Meredith

Database and a decision tree analysis is first used to

develop a customized universe that has an improved

response rate. Then a regression model is built to

determine which variables are the best predictors for

identifying responders. If supplied, Meredith will also

incorporate donation or order amount into the modeling

process in order effectively maximize both performance

measurements. A scoring equation is developed which is

used to select names for roll out. A detailed model

report, including model variables, gains chart, and

ordering recommendations, will be provided upon

completion of the build. Turnaround time for model

build: 4-6 weeks

3- Mailed (Response) Regression Model

Send a file of 100M best customers to Meredith. Meredith will use

decision tree analysis to create a customized universe for the initial

cross-section order. Enough names need to be mailed to yield 800 -

1,000 paid responders. Meredith saves the analysis data on your

cross section at the time names are selected. Once response is

complete (or payment if building net paid model), send a file of the

responders and a copy of the net names mailed file to Meredith.

Once the file is received, Meredith will match the mail tape and the

responders to the saved analysis data and use regression analysis to

determine which variables are the best predictors for identifying

responders. A scoring equation is developed which is used to select

names for roll out. A detailed model report, including model

variables, gains chart, and ordering recommendations, will be

provided upon completion of the build. Turnaround time for

cross-section build: 2-3 weeks. Model build: 4 weeks

4- Decision Tree Analysis

Decision tree analysis can be used alone to create

targeted straight selects or a universe for future

regression modeling. Send a file of 100M best customers

to Meredith to match vs the Database. Using the best

customer matches and a sample of non-matches from

the Meredith Database, decision tree analysis identifies

data points that increase the likelihood of being a best

customer. As each data point is added to the tree, the

resulting segment of customers more closely resemble

the sample of best customers provided. The final result is

a refined universe of Meredith customers that are very

similar to your best customers. The decision tree analysis

provides flexibility in how many names will be available

for future roll-outs. A memo describing the data points in

the refined universe, estimated quantity and lift in match

rate are provided. Turnaround time for model build: