Preparing for Successful VDP

If you are working with direct mail marketing, e-mail marketing (with or without PURLs), mobile media marketing, transpromo or collateral fulfillment and you are trying to personalize the material to improve the response, then you know it is not easy to pull it all together.

The two most difficult parts of working with variable data (regardless of the delivery medium) are working with data to make sure it is accurate and relevant, and then integrating the data into the design of the piece in a meaningful, measurable format. Realistically, one is not more important than the other, because if either one is not properly implemented, the project is likely to fail. So let's start with data.

When people first think of cleaning and preparing a good database they commonly look at the address information, but it is much more than just running the information through a mail prep software package to make sure the addresses are formatted according to United States Postal Service (USPS) standards. Doing this helps to ensure the address is formatted properly by working with the "Coding Accuracy Support System" (CASS). What comes out of CASS is a properly formatted address that is then compared to the USPS's list of actual deliverable addresses. Through the use of "Delivery Point Verification" (DPV) the printer can make sure the address is a real existing business or residence.

Neither CASS nor DPV does anything to verify that the person the mail piece is addressed to is still at that address. For residences you can run the address through the "National Change of Address" (NCOA) service, which looks at who lives at that address and validates the name to the address, but it does not work as well with businesses. A business can complete a change of address card for the USPS if the company moves to a new location, but an individual cannot fill out a change of address card if they change jobs and start working for another business. So you can lose a lot of these contacts in the business world.

What can you do to fix these addresses? Maybe the bigger question is why should you be concerned about fixing them? Let's say you sent out a 20,000-piece mailing and 5 percent of your addresses were deemed undeliverable, and of the pieces that were delivered there was a 4 percent response rate, which generated an average sale of $200. The lost revenue from these bad addresses would be $8,000 (i.e., 20,000 pieces x .05 undeliverable addresses x.04 response rate x $200 per sale). And that does not consider the lifetime value of a customer.

Is it worth spending $100, $200 and $500 to bring in that additional revenue? If the same mailing went to 200,0000 people, the potential revenue would be $80,000. At what point is it worth it to fix the bad addresses that did not make it through CASS/DPV or NCOA? It obviously depends on the size of the job, the value of the response, the nature of repeat orders from the customers, the cost of the piece and the cost of mailing.

Fixing Undeliverable Addresses

What can you do to fix these undeliverable addresses? To start, the USPS has an additional service called Address Element Correction I (AEC I) and Address Element Correction II (AEC II). AEC I is a service offered by the USPS (through one of the mail prep software manufacturers) at $15 per thousand. It has an average correction rate of 30.5 percent. The AEC II is also offered by the USPS to correct addresses and actually gets the local postal carrier involved in the address resolution. The correction rate for the AEC II is 75 percent.

That still typically leaves 25 percent of the undeliverable mail that has not been corrected. These could be people that moved to a new address more than four years ago and are not found in the USPS database records, or maybe they moved out of the country.

So what types of alternatives are available to go after the last group? You can upload the list to a list service provider (e.g. USAData, Info–USA, AccuData, Melissia Data, List–Bazaar, DataFlux, etc.) and ask them to run these names through a comparison to their data–base. Most of these companies keep track of virtually every address a person has had since they were living on their own. The charge is typically per thousand, and it is possible to group multiple databases as one, to contract for a lower price on the overall group rather than pricing a separate list of 100 names.

At this point you are only sending names to this list provider that failed as they were passed through CASS, DPV, NCOA, AEC I, and ACE II; you should be down to somewhere between 1 to 2 percent of the original database that are still undeliverable. In the earlier example of 20,000 starting records, there would be approximately 250 records that would need to be uploaded for correction (20,000 x .05 percent undeliverable x .25 of undeliverable not able to be fixed in AEC I and II). Most of these companies will also run the list through the deceased record list. In addition to this there could be people that have moved out of the country or left no forwarding address with anyone.

Figure 1 shows what a list provider tracks. This sample was taken from a Web service that allows you to search one record at a time for $.95. You could get unlimited searches for 24 hours for $9.95. If you only need 10, 15 or 50, this may be the most cost-effective solution.

Now that the data is cleaned up as far as the addresses, it is time to bring data together to create more relevance for the customer. In addition to the data on the original data list, you can create relevance two different ways.

First, the same list providers mentioned previously will take the list of names you have and append data they are tracking on these people gathered from census information, existing mail order buyers/subscribers, real estate records (deeds/tax assessors), voter registration, magazine subscriptions, survey responses, etc. The data can be demographic facts, psychographics (the use of demographics to study and measure attitudes, values, lifestyles and opinions for marketing purposes) or behavioral characteristics with respect to the customer.

The second option is to create a relational data–base by linking relevant data to a specific field from the original database, adding data to the personalization of the known data from the original database. For example, if you know what year someone graduated from high school or college, you can track significant events/dates/accomplishments that happened during that year. If you know the home state where someone was born, you can link to the state bird, tree, flower, etc. This could even be specific to a given topic like the type of pet they have and related facts to that breed of dog or cat.

Most grocery stores now use a discount card for weekly specials. They have a tremendous amount of information on the customer's purchasing behavior. Knowing one or two specific pieces of information about an individual or company, it becomes possible to create a large number of relevant text and images in the variable data piece.

When you think about managing the data for a variable data job, 500 or 1,000 records are pretty easy to manage and correct problems that you might find. What if there are 50,000 records or 100,000 records though? You have to learn how to work in the database and clean up typical mistakes that would not be corrected (it might not even pass CASS) in the mailing programs—things like a trailing space after a name or address, missing gender indicator, city, state or ZIP code, misspelled words (steet or aveneu instead of street or avenue), improper state abbreviation and inconsistent filling of fields (e.g. some people input the P.O. Box in the first address field and the street address in the second; others transpose these items).

Additionally, you can work with complex calculations in the database applications to calculate someone's age from a birthday or someone's anniversary from the date they got married. Realistically, the creative talents of the people managing the data are the only limit to the options—there are hundreds of complex calculations available on the Web, and you can download scripts (e.g. a mini computer program of the procedural steps put into a loop to remove all of the trailing spaces after the first name).

Wanted: A Database Champion

Success is only going to happen when someone in the company becomes a database champion or when the company recognizes the need to hire a database manager. One simple mistake can ruin the entire production piece and open the printer up to significant financial liability.

Two last things need to be mentioned about working with data for your customers. First, you have to have a secure system to protect their data, and make sure you have backup files if something happens to your front-end system. This most likely is going to involve hardware, software, your own physical environment and protocol you establish in your company to limit access to only certified employees.

The second point is that there are various standards, regulations or laws you need to become knowledgeable about if you are doing mailings, e-mails or messaging with mobile media. It goes further than just the process. There are regulations that have to be addressed when you are dealing with certain vertical markets, such as the financial market or the pharmaceutical market. It is not a regulation or law, but you might want to look into Errors and Omissions Insurance.

Again the design of the piece to create a relevant message is just as critical as the data. The information covered here is just a starting point. Look for Part II of Preparing for Successful VDP in an upcoming issue of IPG, when we will focus on design issues. IPG

John Leininger is a professor in the Department of Graphic Communications at Clemson University. He has been at Clemson since 1986. He has taught courses in flexography, lithography, digital printing, inks and substrates, as well as the department's management class dealing with estimating, planning, equipment purchasing, cost analysis and plant layout. Currently he is focused heavily on the digital printing and variable data market. Contact him at: ljohn@clemson.edu