Presentation - DART Marketing
Transcript Presentation - DART Marketing
Dart Marketing, LLC
Company Introduction & Services
We Provide Solutions to Your Toughest Challenges…
Pricing Strategy - Identify price elastic products and customers
Brand Mapping for better positioning, up-selling and new opportunities
Custom Decision Making Tools for "what if" scenarios
"Natural Selection" Recommendation System
OUR MISSION is to provide insightful and financially responsible marketing strategies, based on
solid experience, sound market data, and state of the art statistical techniques and data analysis.
OUR APPROACH is data driven and our results are measurable. We will better define and target
your most profitable customers, increase your response rates, and improve your marketing
OUR EXPERIENCE includes work with leading publishers (such as Hachette Filipacchi Media),
loyalty programs (such as General Motors lease loyalty), subscription services (such as Gevalia
Kaffe), and the largest direct mail new product launch ever (the GM Card).
OUR GOAL is to return multiples of your cost in increased sales and efficiency.
Services / Capabilities
Analysis, design, and management
Data mining, data quality and data conversion
Customer Relationships (CRM)
Customer profiles and segmentation
Lifetime value of a customer
Market sizing and definition
Marketing Mix Allocation
Sales forecasting and tracking
Determine price elasticities
Price testing and strategy
Time series analysis
Analyze sales patterns, markets and behavior
Perceptual mapping - product & customer matching
Financial modeling - "what-if" scenario testing
ROI and NPV of cash flows
Business case development
Customer behavior, program evaluations, consumer studies
Market sizing, product pricing and configuration, customer
Test design, sample size and definition, questionnaires and
Targeting / Modeling
Maximize objective using regression (i.e. response, incremental
sales, marketing mix, product pricing)
Segmentation - cluster and factor analysis
Custom Analytic Software
”Point-and-click" marketing and budgeting tools
Strategic decision making tools
New business presentation models
Website development, promotion, and strategic assessment
DM creative strategy, copy, design, and consultation
Specializing in publishing for the last 20 years
Overview: Customer and employee relationships (CRM) - Customer profiles and segmentation, lifetime
value of a customer, market sizing and definition, market and customer behavior, time series analysis,
response forecasting and tracking, analyze sales patterns, perceptual mapping.
Customer Profiles and Segmentation
Compare mutually exclusive audience types (buyers versus non buyers, heavy users versus light users,
and employees versus non employees) and product types (premium versus value or newsstand versus
subscription) by demographics (i.e. age, income) and other information such as purchase and payment
history. Results are tested for statistical significance to identify actionable targets.
Lifetime Value of a Customer
Determined lifetime value for GM Card accounts. The value differed widely by the method they were
booked (i.e. direct mail, outbound phones or "take one" applications). The value also differed widely by
customer segment. These findings led them to avoid acquiring accounts through outbound
telemarketing and to increase retention spending on select people.
Price Elasticities, Testing & Strategy
Identify which products and people to raise (and lower) prices to scientifically. Follow with price testing,
coordinated with prices from other channels, such as online and in retail outlets.
Market Sizing and Definition
Sized the market for a new drug for Bristol-Myers Squibb based on population demographics, mortality
rates, and treated and untreated universe. Markets were defined by the disease for which the drug
could be applied.
Market and Customer Behavior
Presented the results of Gold Card analysis to GM Card senior management before its launch,
demonstrating its impact on the number of accounts and total profits under different rewards scenarios.
Response Forecasting and Tracking Models
Ability to forecast and track response to your direct mail. Built accurate response forecasting and tracking
models for mail waves ranging from 15 to 30 million pieces. The GM Card relied on these forecasts to
execute the next mail wave within a tight window of opportunity, maximizing response.
Time Series Analysis
Track sales over time, accounting for seasonality. Pre and post marketing event analysis. Example:
Showed changes in response to various methods of account acquisition over time by demographics,
credit variables and the vehicle ownership of those solicited for a GM credit card. Also used for tracking
and forecasting and optimizing newsstand sales and returns.
Analyze Sales Patterns
Identify seasonality and region brand strength by comparing your sales to your competitors. Example:
Measured sale rates for Cadillac brands versus competitors by city to identify regions of brand strength
and weakness for direct mail campaign.
Perceptual Mapping / Competitive Perceptual Maps
3-D bubble plots illustrate how products correlate with each other on various dimensions, such as by
sales correlations, price/offer and demographics. The size of each bubble represents volume, such as the
number of subscribers/sales or revenue. See an example on the next page.
Taken a step further, we can create an individualized product selection model to determine which product
each customer will most likely want next.
Personalized Recommendation Systems
According to ValueBeat, Amazon.com attributes 35% of their product sales to recommendations.
We create a truly personal shopping experience with My-Store, WishLists, email alerts, smart-search
System becomes your virtual aisle clerk, assisting customers at time of decision-making
Simple to integrate and automatically tunes its results and runs in real time without any investment in
infrastructure, training or maintenance.
We provide personalized recommendations that…
Are relevant. Match customers tastes. Lead to discovery of new products. And, engage the customer.
We then customize it to…
Leverage profit margin or price. Remove out of stock merchandise. Promote sale items or overstock. Make
layered recommendations, such as best “in category” and “out of category”. Be controlled by the user to
guide recommendations, even for New Shoppers. Treat customers differently based on tastes.
Compare our movie recommendations to other world-class systems. Try Our Demo
Nat Lamp.'s Animal House
Nat Lamp.'s Vacation
Nat Lamp.'s Animal House
Nat Lamp.'s Vacation
Nat. Lamp.'s Van Wilder
Nat. Lamp.'s Christmas Vacation
Notes: Results taken from Netflix.com, Blockbuster.com, and FamilyVideo.com in November 2006.
Financial modeling projects range from a single mail campaign to a multi-year cash flow projection
of your entire business. It brings together assumptions about the product, customers, budgets,
and marketing effectiveness using historical data, market research and experience.
We will help you establish a base case and enable you to test marketing investment trade-offs,
identify high leverage opportunities, find break-even points, measure performance against plan
and test “what-ifs.”
The model tracks a prospect from solicitation to sale to customer retention.
Modeling allows you to evaluate the big picture, with the ability to understand differences in
performance segmented by market and product.
The model evolves over time, becoming increasingly accurate and relevant, as it is continuously
updated with future program results and further customized.
It serves as an excellent custom reporting device (i.e. automated scorecard).
GM Card: Provided business analysis for the GM Card program in the U.S., as well as subsequent
models for launches in Canada, Mexico, England and Brazil. The models provide a complete picture of
the company, incorporating assumptions about the product, economic factors, banking fees, marketing
budgets and the effectiveness of each marketing channel.
As a result of these models, GM confidently expanded into Canada and England, cautiously entered
Brazil and avoided Mexico. Different product configurations were offered in each country based on
maximizing the results from each financial model.
After the launch, the U.S. model was updated to include actual data and drove decision making for ten
years, including the decisions to:
Discontinue the partner program
Eliminate to the gold card product
Shift to a new tiered rewards program
Bristol-Myers Squibb: Our custom financial and marketing planning tool sized the market for a new
drug based on population demographics, mortality rates, and treated and untreated universe by
disease for which the drug can be marketed. It incorporated details about media spending, DTC,
professional spending, and a compliance and persistency program. The user could then test various
marketing strategies and budgets, and learn the expected sales mix and financial results.
Product Perceptual Mapping
There were 16 products classified into 5 product
segments. The products were plotted on three
dimensions, where the x and y axis represent
sales correlations between the products and the
z axis represents unit sales volume (sometime
revenue is used instead) shown by the size of
the bubble. Demand appears to be higher on
the left part of the map.
The products of each segment tended to form
mini-clusters and larger "segment clusters"
tended to form as well, giving both a macro and
micro view on one page.
Key - Baseball Card Sets
Great World Series Upsets
MLB 2000 Season Cards
Best Teams Ever
Subway Series Set
Best Players of the Live Ball Era
Best Players of the Post War Era
Great World Series Matchups
Best Players of the Modern Era
MLB 2001 Season Cards
Best Players of the Dead Ball Era
Teams of the 1800's
MLB 2002 Season Cards
Japanese Franchise All-Stars
MLB 2003 Season Cards
Prime Time Franchise All-Stars
• The horizontal axis is a scale of how “exotic” the product is
• The vertical axis appears to be a scale of real life “best of.”
• The more exotic the product the more pricing power
Real Life <-------> "Best of"
This plot shows sales correlations,
between baseball card sets.
The scales of these maps are highly interpretive.
HIGH PRICING POWER
Team Sets T-WS
SCALE OF EXOTICNESS
Real Teams <-------- Team Collections ---------> Player Sets
Competitive Perceptual Maps
Illustrate how products correlate with each other on various dimensions.
Types of Maps:
If we focused only on men’s
magazines, the gender
dimension would be replaced
with something more salient,
such as price.
More sophisticated plots, which
require advanced statistical
analysis, are even more
compelling because they show
sales correlations, based on
past acquisition mailings or
In this simple example, a broad
array of magazines were
plotted by their similarity to
each other on two dimensions,
median age and gender. The
size of each bubble represents
the number of subscribers.
Track entire marketing effort including all budgets and expected results for each media, compare
results to benchmarks and link to a financial model. Example: Pre launch research and marketing
plan for the Chrysler Credit Card. The model demonstrates how any change in the marketing
assumptions impacts the bottom line.
Analytic Marketing Tools / Scenario Testing
We build custom "point-and-click" analytic marketing tools using Excel and Visual Basic. Example:
Showed changes in response to various methods of account acquisition over time by
demographics, credit variables and the vehicle ownership of those solicited for a GM credit card.
This analysis identified highly responsive targets, making the GM Card's marketing team more
efficient in acquiring millions of new accounts. Another model tested different product
configurations to determine their pricing strategy.
User-Friendly Front Ends to Databases and Statistical Forecasts
Example: Built easy to use software enabling the GM Card marketing team to access revenue and
redemption expense forecasts in graphical and table form, which were tied to a historical database.
New Business Presentations
Create "point and click" models to demonstrate the proposed marketing plan and its effectiveness
on sales and ROI. Example: Bristol-Myers Squibb Ortho birth control pills.
Ask to see demos of these types of solutions.
Overview: We research new products and services, do program evaluations, understand customer
behavior, do consumer studies, and do on-the-fly cover testing. We use everything from phone surveys
to mall intercepts to online research.
We will help you identify relevant groups to study or compare. If you are looking at buyer behavior for
example, look at non buyer behavior or have a control group or program norms as a reference. We may
collect surveys from people different each time periods to remove seasonality or create sales weights.
New Product Research
New product pre-launch research. Example: Safe-card's child protection program tested different prices
and different offer combinations. It also established target segmentation for each service.
Measure program effectiveness and achieve higher sales rates. Example: Found insights into consumer
behavior for GM's North American automotive marketing team, enabling them to achieve higher (more
than double in some cases) incremental sales rates in subsequent programs.
Determine best motivational factors for each target group. Example: Measured GM employee attitudes
and willingness to participate GM’s Ambassador Program for hourly, salaried and retired employees.
On-the-Fly Cover Testing
Get quantifiable consumer reactions to magazine covers before choosing which to use. Measure lift of
cover choice to sales by calibrating to survey responses. Predict the overall success of the issue before
release, based on live consumer feedback.
Overview: Customer Relationship Management is at the core of database analysis and purpose. Our
methods follow...database design and management, data mining, data quality, data hygiene,
merge/purge/de-dupe, and data conversion.
Experience with files containing millions of records. Examples: GM sales files, GM Credit Card customer
files, and various other transactions level data.
Reorganized Prodigy Internet's customer database under intense time pressure to produce marketing
recommendations for new acquisitions, by identifying heavy users of their service and their
Is your data a mess? Try this one on for size. GM had a direct mail program, where people responded
by hand writing in the brand they were interested in buying. Many of these were misspelled or invalid
and we wrote an algorhythn, which identified the brands with over 98% accuracy.
Do you have lists that need to be combined or reworked? Are you mailing twice to the same address?
Do you want to mail only to certain segments of your list? Need help identifying who those people
should be? No list is too big for us to manage.
Translate files from one file format into another, such as from ASCII fixed column to a SAS dataset.
Overview: Maximize response, incremental sales and marketing mix, forecasting and tracking,and
Maximize Objective Using Various Regression Techniques – Such as response rates for
acquisitions, renewals, incremental sales, customer targeting, and marketing mix. Example: built
models for GM Card identifying prospects most likely to not only respond to an offer, but to pass the
credit check and become a booked account.
Target Models – identify affinity points programs redeemers, likely program defectors, product
switchers & migrators within program, brand preference, the timing of market entry, and heavy users.
Example: built a GM Card redeemer model identifying the likelihood each customer would redeem
their rewards and how soon for millions of cardholders.
Profiling and Segmentation - identify audience types such as defectors, in market buyers, ethnic
groups, military/college buyers, employees, owners/loyalty, competitors/conquest, and unique
segmentation schemes using multivariate analysis. Example: Segmented GM’s hourly, salaried and
retired employees into research defined “attitude and energy” dimensions for inclusion in their
Ambassador Program, designed to encourage and empower employees to sell cars.
Forecasting and Tracking – forecasting; including seasonality and other patterns using previous
months or years of data and assumptions about future trends. Results are tracked and compared to
forecasts. Example: account and revenue forecasts for the GM Card.
Significance Testing - Distinguish between results that are directionally and statistically significant.
About Craig Tomarkin
Craig Tomarkin founded DART Marketing with the vision of providing intelligent, financially responsible
marketing strategies through custom analysis. His team is well suited for any analytic challenge.
After completing business school in 1990, Craig joined InterData (a privately held spin-off
of DMB&B's direct marketing group), where he immediately became a key member of the
GM Card launch team. His sophisticated financial models were used to establish the
expected business case and allowed them to play what-if scenarios on pricing, product
configuration, and rewards payments. In the first year of business, the program booked
five million accounts, the most successful launch of its kind.
After the launch, he updated the U.S. model to include actual data and drove decision making for ten years,
including the decisions to:
- Discontinue the partner program
- Eliminate the gold card product
- Shift to a new tiered rewards product
Other Highlights Include:
Improving Hachette Filipacchi's magazine renewals using regression models combined with an innovative
mail strategy. Titles modeled include: Woman's Day, Home, Road & Track and others.
Unique product perceptual mapping of Gevalia's coffees. Brands clustered into Flavored, Roasts, and Exotic.
When price was later added as a factor, the more expensive coffees moved up the exoticness scale. This
had implictions for pricing strategy, new product positioning, and targeted up-selling.
Also to his credit, Craig invented the “Natural Selection” Product Recommendation System and founded
Baseballguru.com, a popular baseball web site that continues to grow in popularity and profits.
Apply the science of
Learn more about your
customers and improve
your bottom line.
Contact us today!
DART Marketing, LLC
2333 Congress St.
Fairfield, CT 06824