×
DMA: Data and Marketing Association
Why DMA?
marketing-analytics

Hands-on Predictive Analytics for Marketers

Learn to build your own basic “quick and dirty” model that leverages multiple variables to predict various marketing outcomes of interest.

DATE
TIME
LOCATION
June 19, 2017
9:00am - 5:00pm
New York, NY
Register
December 13, 2017
9:00am - 5:00pm
New York, NY
Register

Overview

Predictive Modeling is both an art and a science that requires considerable training, investment, and resources. As a result, marketers with limited budget are often not able to fully leverage the power of predictive modeling. This course is designed for the marketer that seeks to quickly build predictive models with limited resources, short time frames, and manageable data sets. The course includes hands-on instruction on the following components of predictive modeling.

Topics

  • Selecting the Marketing Variable that needs to be Predicted
  • Building an Analysis Data Set
  • Data Preparation
  • Building Model Algorithms
  • Validating and Evaluating Models
  • Using Models to make Predictions
  • Profiling Data based on Models

About The Instructor

headshot
Devyani Sadh
As CEO at Data Square, Dr. Sadh, has successfully spearheaded numerous programs for CRM and Data Infrastructures, Hygiene, BI, Predictive Analytics, Digital ... Read More »

Course Pricing

  • dollar Member - $799
  • dollar Non-Member - $1,399

Early Bird Discount! Register up to 8 weeks prior to the start date for 20% off the above rates. Price reflected at checkout.

See The DMA Member Benefits »

Course Duration

  • 1 day

Selecting the Marketing Variable that needs to be Predicted

  • Defining the marketing objective
  • Specifying the dependent variable

Building an Analysis Dataset

  • Sampling techniques
  • Splitting the data for model development vs. validations

Data Preparation

  • Examining variable distributions
  • Treatment of continuous vs. categorical variables
  • Transforming variables
  • Implementing methods of imputing missing values
  • Outlier detection and treatment

Building a Model Algorithm

  • Understanding and selecting key predictors
  • Combing multiple predictors to develop a model algorithm
  • Selecting between models

Validating and Evaluating Models

  • Assessing Predictor Variable Impact
  • Understanding Model Goodness-of-fit measures
  • Creating a Gains chart to evaluate model performance

Using Models to make Predictions

  • Scoring a new data set on the model to make predictions
  • Creating model deciles to group data
  • Examining distribution differences between model development and scoring universe groups

Profiling based on Model Results

  • Building single-variable descriptive profiles based on model groupings

 

Tips and tricks to simplify complex data issues will also be discussed. By the end of the day, you will have the skills to build your own basic “quick and dirty” model that leverages multiple variables to predict various marketing outcomes of interest such as response rate and purchase rate. You will also be able to assess the utility of your model to meet your marketing objectives. Future parts within the series will cover more sophisticated processes and methods to further enhnace the accuracy of the modeling methodology as well as to deal with common marketing data challenges.

*Outline is subject to change

Are you just dying to attend one of DMA’s Education seminars this year, but you’re not sure how to convince your boss to let you attend? Try one of these approaches.

Try:
  • “I know budgets are tight, but this is an opportunity that pays for itself instantly. A single tip from any of the DMA seminar could pay for the costs of the whole training. I’ll be learning secrets from practicing experts and industry thought leaders.”
  • “I’ll become more proficient in industry best practices, as well as on the latest tactics and technologies in today’s ever‐evolving marketing landscape. From direct mail to data governance, social media to email marketing; the DMA has all of our needs covered.”

I can share what I learned after the seminar; maybe host a brown‐bag working lunch and present the team with all the new techniques, strategies, and ideas I’ll come back with. Combining my own notes and ideas with the session handouts provided by the DMA, I’ll have plenty of inspiration and information to share when I return. Besides… this’ll allow everyone to get trained all for the price of sending just one person; and the team will definitely appreciate the opportunity to find out about the latest strategies used by many of our competitors today.

Still presenting itself as a hard sell? Inform your boss that your course can go on their accomplishments at the end of the year for their review: “Trained staff to improve results!” This is a win/win.

Want to do even more with this topic? Join our Analytics Community and participate in knowledge sharing, best practice development, leadership building and networking with DMA members from across the globe. Please note, you must be a DMA member to participate in this group.
Learn More
Our Brands EducationEventsAdvocacy MembershipAccountability ResourcesKnowledge CenterAbout UsBlogContact Us

Login To Your Account

Send this to friend