Subscription Services

1
Subscription Services Get more customers Get
more business from existing customers Make your
marketing more efficient and effective
2
Contents

• Subscription Acquisitions and Renewals3
• Response Modeling.4
• Financials and Profiles5
• Test Case Example6
• Where Do We Go from Here12
• Optimizing Your Mail and Price Strategy.13
• Improving Response Model Effectiveness14
• Test Improving Response through Research and
  Segmentation15

3

• Subscription Acquisitions and Renewals
• Response Modeling
• Financials Profiles

4
Response Modeling

• DART builds sophisticated custom models that
  generate superior results.
• Purpose
• To make your direct mail and telemarketing more
  efficient and effective.
• Method
• Our highly experienced modelers predict
  responders with remarkable precision. They
  achieve this by scientifically determining the
  degree to which all of the data contribute to the
  likelihood of response. Models can also be built
  with other objectives in mind, such as maximizing
  revenue or predicting price sensitivity. Using
  which ever form of regression proves most
  effective, variables are reduced to include only
  those contributing to the solution. Data quality
  checks are performed and any problems or
  questions are reported before the model is
  complete.
• The standard write-up includes a description of
  the model and a gains chart.
• The expanded write-up adds 1) Descriptive
  statistics - a basic report summarizing what data
  was used and what the mail file data
  distributions were. 2) Descriptive profiles -
  shows responders versus non-responders. 3) Decile
  Analysis - compares the top three deciles to the
  bottom three for more distinctive profiles. 4) A
  one-page model evaluation and validation summary.
  5) A two-page financial analysis showing expected
  performance by decile and in summary.
• Results
• An elegant model that makes sense, will be easy
  to implement, and will greatly reduce mailing
  costs.

5
Expanded Write-up Financials and
Profiles

• We provide a detailed analysis and an expanded
  write-up, which makes sense out of the numbers.
  Each of these can be provided a la carte as well.
• Descriptive Statistics
• A basic report summarizing what data was used and
  what the mail file data distributions were.
• Descriptive Profiles
• This analysis describes the attributes of
  responders. For example, Responders are 36 more
  likely to live in large cities than those in the
  mail file. This also reveals targets for your
  creative.
• Decile Analysis
• This analysis profiles the attributes of those
  the model identifies as likely to respond. It
  compares subscribers in the top three deciles to
  those in the bottom three deciles.
• Model Evaluation and Validation
• This is the final step in the analysis process.
  We compare actual responders to modeled
  responders to validate the model and screen for
  inconsistencies.
• Financial Decile Analysis
• This analysis shows the expected financial
  performance of each decile. The analysis is
  thorough, considering everything from response
  rates to the residual revenue from expected
  future sales.

6

• Subscription Response Modeling
• Test Case Example

7
Decile Table of Model Performance (Gains Chart)

• Test Case This is an acquisitions response model
  example (we also build pricing, revenue and
  renewal models). Those mailed in the top three
  deciles were more than five times as likely to
  subscribe as those in the bottom three deciles.
  To provide added precision, we can also divide
  these into ventiles (20 groups) or simply use a
  raw score to allow selection of the desired
  number of names.

8
Descriptive Profiles, Decile Analysis and Model
Validation

• Description
• This table compare profiles of actual responders
  to the modeled responders.
• The Descriptive Profiles are on the left, the
  Decile Analysis is on the right.
• The full version of this write-up includes
  charts. The Decile Analysis also includes brief
  descriptions identifying statistically
  significant findings.
• Model Validation of the Test Case
• The similarity in the indexes shows that the
  model is doing an excellent job of assigning
  responders to the top three deciles for each
  variable.

9
Financial Decile Analysis
10
Financial Summary

• Description
• This table shows the expected results from one
  mail effort.
• Key Findings
• Mailing only the top half of the list generates
• 83 of the responders,
• 86 of all revenue, and
• 98 of residual revenue.
• Given a mail file of 747k names, mailing only the
  top half saves 187k per mail effort.

11

• Where Do We Go from Here?
• Optimizing your mail strategy
• Increasing prices without reducing sales
• Improving response model effectiveness
• Improving response through research and
  segmentation

12
Where do we go from here?

• Get more subscriptions at higher prices for less
  money using services designed to pay for
  themselves through the value they return.
• Test, monitor, and evaluate
• We guide you through this process to ensure
  consistent improvement in direct marketing
  results.
• Optimize your pricing strategy
• Statistically determine who is willing to pay
  what price. Once the results from the mailing are
  in (about 3-4 weeks worth of responses are
  adequate), we can fine tune pricing strategy by
  decile. Requires structured prices tests during
  mailings.
• Optimize your mail strategy
• Financial modeling increases precision by
  measuring the affect of offers and mailing
  different audiences the right number of times.
  Set your plan and test what-if scenarios.
• Improve the effectiveness of the response model
• Database Enhancement Research with Cost/Benefit
  Analysis
• Other ways to improve response rates
• Custom Segmentation I am not the median.
• Pre Mail Research Understanding purchase
  motivation and behavior, and desirable product
  attributes leads to better targeted creative and
  better response.
• Post Mail Research Understanding the why
  leads to more effective offers.

13
How do we optimize our mail strategy and pricing
strategy?

• Pricing Analysis and Financial Modeling
• Purpose
• Sell more magazines, at higher prices, for the
  least amount of money.
• Method
• We statistically determine who we can raise
  prices to, by how much, and on which products.
• Follow up with a custom financial model, further
  customized to consider timing and list fatigue
  from multiple mail efforts, different packages
  and prices. Predict with relative certainty the
  results of the mailing.
• Audiences and Packages - Allow for mailing only
  cherry picked audiences and packages within a
  decile, factoring in segmented performance.
• Timing and List Fatigue - Allow what-if scenarios
  by each mail effort.
• Multiple Prices - Evaluate the impact of changes
  in price on the fly. For best results, a
  structured price test or a long history of prices
  and response rates is required. The price test
  can be non-invasive and not significantly alter
  the results of any campaign. In absence of a
  price test, a linear price to response index is
  used in its place for each target audience.
• Results
• Expect more responses at higher prices for less
  money.

14
Are we missing data that will improve model
performance cost effectively?

• Database Enhancement Research with Cost/Benefit
  Analysis
• Purpose
• Examine the value of adding additional
  demographic and other variables to your database,
  and evaluate whether the benefits from increased
  model performance out weight the costs.
• Method
• Establish baseline - start with a model you
  already have (or build a new model) and include
  the variables currently available on your
  database.
• Test additional data - append a large number of
  variables from a company like Experian or
  Donnelley to a small sample of your customer and
  mail files to minimize data costs.
• Identify cost effective contributing variables -
  examine the new variables to see if they add
  efficiency to the model and include only those
  that make economic sense to add (either due to
  decreased mail cost or increased response).
• Results
• The value in this analysis is particularly
  evident long term. A residual benefit is any new
  information that leads to better targeting your
  creative.

15
How can we improve response rates among those we
finally choose to mail?

• Custom Segmentation I am not the median.
• Purpose
• Identify distinct target audiences within your
  subscriber base.
• Method
• Segmentation can be achieved either through
  obvious customer groups or through sophisticated
  statistical clustering techniques. Market to each
  segment individually with distinct creative,
  rather than treating everyone as if they were the
  median subscriber.
• This technique is often coupled with pre mail
  market research to deliver highly informative
  creative and increase acceptance rates
  significantly within each target audience.
• Results
• Better targeting leads to better response rates.

16
How can we improve response rates among those we
finally choose to mail?

• Pre Mail Research Understanding purchase
  motivation and behavior, and desirable product
  attributes leads to better targeted creative and
  better response.
• Purpose
• Gain customer intelligence and understand
  purchase motivation by target audience. This
  leads to better targeted messages in your direct
  mail creative, increasing response rates.
• Method
• Through market research, find out what attributes
  of the magazine appeal most to the each target
  audience and which behavioral characteristics
  link to likelihood to buy. It may boil down to
  one or two key attributes that indicate whether
  someone is a good prospect and one or two key or
  behavioral characteristics that indicate why they
  buy.
• Results
• For example, Corvette buyers like to drive fast
  not to drive through winding scenic roads. It
  turns out that need for speed is different from
  need for performance. And, Corvette buyers are
  more likely to be middle aged men nostalgic for
  their childhood dream car, rather than young men
  dreaming of this car.

17
How can we do better next time?

• Post Mail Research - Understanding the why
  leads to more effective offers.
• Purpose
• Companies generally judge whether or not an offer
  worked solely by the response rate to the
  mailing. If the response was low they increase
  the offer, if it was high they do it again or
  reduce the offer. But, to learn why the offer
  did what it did, follow the mail with surveys
  designed to get additional detail about their
  thought process, purchase consideration, and the
  value they placed on the offer.
• Method
• Implement three surveys as follows 1) awareness
  and readership, 2) buyers, 3) non buyers. It
  could be condensed to remove the awareness and
  readership part, but experience indicates that it
  is worthwhile.
• Results
• This will lead to a better offer next time, that
  will not be more expensive, just more effective.

18
Getting Started
Contact us today! Learn more about how we can
help you achieve better response rates for less
money.

• Craig Tomarkin
• DART Marketing, LLC
• 2333 Congress St.
• Fairfield, CT 06824
• CTomarkin_at_dartm.net
• 203-259-0676
• Fax [phone]