Data Mining, Enhancement and Modeling
Description of Our Data Mining, Enhancement and Modeling
One might ask, what does a communication firm have to do with Data Mining, Enhancement, and Modeling? Isn’t that just for computer techs and statistical geeks? The answer can be inferred from a quote by one of our favorite philosophers, Aristotle, who said in his famous treatise On Rhetoric, “Rhetoric [by that he meant persuasion]is the art of discovering in the particular case all of the available means of persuasion.”
That is a perfect description of how we view the relationship between Data Mining, Enhancement and Modeling — indeed all marketing research — and effective communication. The database contains (or we can construct it to contain) all of the available means of persuasion for each particular constituent in a particular category (e.g., prospect, customer, donor or voter). The mining discovers the relative persuasive power of each of those available means (variables.) The modeling develops a formula that predicts the relative persuasive effect of variables when combined.
When viewed from this perspective, especially in today’s environment of one-on-one direct marketing, Data Mining, Enhancement and Modeling is the core domain of an effective communication firm. With over 30 years of experience in Data Mining and Database Modeling, Steven Ethridge knows how to use Data Mining, Enhancement and Modeling to increase your Marketing/Communication Return on Investment.
Unique Features of Our Data Mining, Enhancement and Modeling
Several qualities make our services in this area unique, as follows.
- Cost Savings — Most direct marketing firms go to the expense of creating final productions, printing hundreds of quantities and mailing them to matched samples of people to see which concept performs best before mailing to the entire target database. Then they decide which concept generated the highest response, build a model to identify the demographic characteristics of the people who responded to that concept, and select a larger group of people like them to mail that concept to again. We save clients money by testing messages through surveys and appending survey data to the database to build the model rather than testing costly mailings. In today’s world of web survey capabilities, it is no longer sensible to spend money printing and mailing hundreds of pieces to test direct mail concepts.
- Advanced Database Modeling — Our second unique feature of our Data Mining and Database Modeling is our experience in a wide range of advanced, multivariate statistical modeling techniques. These include Linear Regression, Logistic Regression, Discriminant Analysis, (Chi Square Automatic Interaction Detector (CHAID), Neural Networks and more (for a complete list of our statistical capabilities see our Data Analysis section of our Research page.
- Analysis of Unstructured Data — Third is our experience working with unstructured data. We have data mining tools that allow us to capture, code and analyze virtually any kind of digital data, whether transcripts of verbatim responses, print, pictures, audio, or video, and analyze all of these as inputs to the model.
- Merge-Purge Services — If you have several different customer lists, each of which is in a different file layout or format, we can merge those into a single file with a standard layout, purge out duplicate records and return to you one single file.
- List Brokering — We are qualified List Brokers for many of the largest and best- respected direct mail marketing list firms in the world. This means that we can find and append a wide range of demographic and psychographic data variables to your database that can be used for market segmentation and modeling.
- Wealth Screening — Used primarily for direct mail fundraising database development, through our list brokering relationships we have the ability to take your donor file and screen it through a database of public information on more than 7 million of the wealthiest individuals in the U.S. to identify individuals who have the greatest potential for giving major gifts to your non-profit.
- Phone Appending — We can take your database of customer, member, donor or voter addresses, append phone numbers to those addresses and, if necessary, scrub out phone numbers that are in federal and state Do Not Call (DNC) lists.
- Email Appending — In addition to list brokering, we can take your list of customers, members, donors or voters, match them against a list of over 500 million opt-in email addresses, and then help you develop Can-Span compliant email marketing campaigns.
- Reverse Email Appending — If all you have is an email address and you need to append the customer’s name, mailing address and telephone number, we can do a reverse email append to add this other contact information to your database.
Key Benefit Deliverables from Our Data Mining, Enhancing and Modeling
We begin by understanding your communication objectives. Then, we examine your database to determine what needs to be done to enhance your file into a “Smart Database” that will help you dramatically increase your Marketing/Communication Return on Investment. At that point, we submit a plan with recommendations that could include any one or a combination of the services listed above. Each database project is custom designed and implemented to your unique needs. Upon completion of the project, the key benefit deliverables typically include one or more of the following.
- A “Smart Database” — Database Architecture to return to you a database that contains intelligent data variables that you can use to target specific types of direct marketing messages to the people who are most likely to respond favorably to those messages
- Database Segmentation — Segmentation of your database on predictive variables and then categorization and rank ordering of people by their likelihood of responding to your specific direct marketing campaign
- Model Variables and Formula — A listing of the variables and the formula that you can use to calculate the probability of response for individuals that you may later add to your database
- Analysis with Recommendations — An Executive Summary Report with Recommendations for direct marketing strategies to increase your Marketing/Communication Return on Investment
Types of Clients That Use Our Data Mining, Enhancing and Modeling Services
Any organization that needs to cost-effectively reach their constituents (customers, prospects, donors, members or voters) using direct marketing (whether direct mail, telemarketing or email marketing), can benefit from our data mining, enhancing and modeling services. Such organizations include banks, credit unions, investment banking firms, hotels, airlines, automobile rental companies, non-profits for fundraising, call centers, political campaigns, referendum campaigns, retail stores, and many others.
A Case Study Demonstrating That Our Data Mining, Enhancement and Modeling Services Work
We recently had a client hire us to help them with a voter registration campaign. The tasks were to (a) discover what messages would persuade the largest percentage of people to vote for their side of the issue and (b) identify, by name and address, which individuals, who are not registered to vote, could most likely be persuaded to their side of the issue; then (c) to register to vote as many of those people as possible. The steps that we took, which can be applied to any direct marketing situation, were as follows:
- We conducted a quantitative survey that identified the key motivations, and tested messages, as well as identified the Persuadable Segment.
- Then, based on the survey data, we developed a statistical model that predicted the probability that each individual in the survey sample would be in the targeted Persuadable Segment.
- Next, we obtained from a reputable marketing list vendor, the list of every adult of voting age in the area and ran his or her data record through the predictive model to identify the individuals that were most likely to be in the client’s target market.
- We then obtained the list of registered voters, merged that list with the list of all adults in the area, and purged out the ones that were already registered to vote.
- Of the remaining list of individuals that were not registered to vote, we then segmented out the ones that the model predicted were likely to be in our target market, the Persuadable Segment.
- We gave the list of individuals that the model predicted to be in the target market to the client’s PR firm. The PR firm then developed a direct mail voter registration campaign that emphasized key messages that we found would work.
- After the campaign, the PR firm obtained an updated list of registered voters to see which ones in our target list had recently registered to vote.
- Whereas the PR firm had typically experienced responses rates of new registered voters around 5%, they found that 51% of our targeted list had registered. This was an unprecedented response rate in the PR firm’s history, which was as a direct result of our modeling.