Direct Mail Marketing – Employing Data Analysis Techniques for Reducing Expenditures

October 3rd, 2010 | Uncategorized | SK | 8 Comments

Many companies attract customers to their product or service offerings via direct mailing marketing.  These mailers are expensive as the average response rate is about 2.61%[1].  For example, to attract 500 responses, companies would have to send the mailers out to 20,000 prospects.  Each mailer sent costs a certain amount of money.  The costs can range from $0.30 to $3 per mailer.  If companies reduce the number of mailers sent, but maintain the same number of responses, it will see significant savings.  Data mining can be used to achieve this reduction of mailers sent.

Benefits of Data Analysis/Data Mining

Data mining helps select customers who have higher probability of responding to the mailers within the target market.  This can be done by assigning a score from, 1 to 100, where 1 means “very unlikely to respond” and 100 means “very likely to respond,” to each potential customer.  The prospects then could be sorted according to this score, and marketing department can then start sending out mailers in batches. Of course, the higher score batches are sent out first.  The marketing department can stop sending out mailers once the target number of responses is met.

The hypothetical response rate for this particular example is shown in the figure above.  It shows that traditional randomized mass mailing will have a response rate much lower than that of the intelligent mass mailing discussed above.  This example shows that the target number of responses would lead to savings of about 5,000 mailings, or about 25%.

Data Analysis

The first step for any data mining analysis is to conduct a qualitative review of the product and interviews with product manager and some prospective customers.  The interviews provide qualitative information that is very useful in creating models, as well as help eliminate uncharacteristic results the models may provide.  This information is critical for selecting an ideal modeling technique.

After selecting a modeling technique, a model is developed such that a score of 1 to 100 is assigned to each potential customer.  This score is based on many factors, including the match between the target market profile of the product and the profile of the customer, which is determined by the customer’s prior response rate to the mailings.  Other complex factors, such as a decision tree or neural network, could be included in the analysis.

A model that is generally used in a direct mailing case is a non-response model.  In a non-response model, prior response rates to the mailings or certain types of mailings will be taken into account in the model.   If the customer has a history of not responding to any mailings or telemarketers’ calls then the customer may get a lower score depending on the model.

[1] DMA survey results, 2003,

8 Responses and Counting...

  • Izabelle Mccullen 10.03.2010

    Thanks-a-mundo for the blog article. Really Great.

  • I find it difficult to identify your rss, just how do i sign up to it?

  • Oh my goodness! an amazing article dude. Thank you However I am experiencing issue with ur rss . Don’t know why Unable to subscribe to it. Is there anyone getting identical rss problem? Anyone who knows kindly respond. Thnkx

  • Hi there! This is my first visit to your blog! We are a team of volunteers and starting a new project in a community in the same niche. Your blog provided us beneficial information to work on. You have done a extraordinary job!

  • I really enjoy the blog post.Thanks Again. Much obliged.

  • Greetings from Idaho! I’m bored to death at work so I decided to browse your website on my iphone during lunch break. I enjoy the info you provide here and can’t wait to take a look when I get home. I’m surprised at how quick your blog loaded on my phone .. I’m not even using WIFI, just 3G .. Anyways, excellent blog!

  • Yet another issue is that video games are normally serious as the name indicated with the key focus on mastering rather than leisure. Although, we have an entertainment feature to keep your young ones engaged, each game is frequently designed to work towards a specific expertise or area, such as math or scientific discipline. Thanks for your posting.

  • Thanks alot : ) for your post. I want to say that the expense of car insurance differs from one scheme to another, since there are so many different facets which give rise to the overall cost. For example, the brand name of the automobile will have an enormous bearing on the price. A reliable older family vehicle will have an inexpensive premium than the usual flashy racecar.