June 13th, 2011 (Germantown, MD) — DATALAB USA (www.datalabusa.com), a 4 time analytical award winning data marketing and predictive modeling company, today announced the completion of its server facility migration project started earlier this year. The Maryland-based company has taken a leap in storage infrastructure using the latest thermal containment and redundant architecture available.

With raised floor/cold aisle structures becoming a thing of the past, DATALAB’s move to hot aisle architecture for data containment is a natural progression for the 32 year old company. The new data center promises to deliver the latest in scalability, modularity and efficiency that can translate to less cost and faster/better response to servicing its client-base.

“Handling our data and servicing our clients go hand-in-hand,” commented Ryder Warehall, EVP of Information Technology and Security. “We have always strived for practicality and our new data center delivers just that –more rack space, higher density racks, and more efficient battery arrays.”

The company is certainly not stagnant when it comes to staying current with the most recent operational trends and best practices. It wasn’t long ago that DATALAB upgraded its backbone connection from multiple DS3 to multiple 100 megabit per second lines, more than doubling its former capacity. These latest infrastructure upgrades underline DATALAB’s dedication to staying competitive as a fully-hosted, self-maintained and controlled data services provider.

About DATALAB USA

A four-time award winning (www.datalabusa.com/awards) data marketing analysis and predictive modeling company, DATALAB USA has been providing top-tier products and services to its clients for over 30 years (www.datalabusa.com/data-solutions). The firm’s specialized industry experience include Fortune 500 companies in the financial, retail, education, insurance and non-profit sectors (www.datalabusa.com/our-clients).

DATALAB USA is a privately held company headquartered just outside of Washington, DC.

For more information, contact Alex Aigner at (800) 972-1430 or aaigner@datalabusa.com

7. Technology with an outsourced partner is constantly being reviewed and updated.  Since so much data is being processed on a daily basis, companies find it necessary to stay on top of the latest technology regarding computers, processing software, modeling software as well as servers and security measures.  Since these vendors are in the data business, technology is considered one of their most precious assets.

8. Experience with various types of audiences:  If you are looking to do it, chances are, a good database marketing vendor has already done it.  For example, isolation of particular demographic populations ranging from interest, income, gender or ethnicity, to education.  They will also likely know the best way to address marketing for a Business-to-Consumer, or Business-to-Business initiative by analyzing the data.

Now that we have reviewed the many advantages, a couple of things that companies sometimes overlook are the skill sets of their internal teams, as well as the cost/benefit ratio.  As discussed earlier, since many companies are all about number crunching, and do not necessarily have the database marketing expertise, some companies, such as DataLab USA (another shameless plug…), are willing to run an analysis on your data sets and create a results simulation based upon prior campaign selections or un-modeled data.  If a company is willing to do a head-to-head test against your internal marketing team, and not charge for the expense, then why pass up the opportunity to investigate the possibilities for improving your campaign performance.

This also leads to the topic of the cost/benefit ratio.  When considering the overhead that organizations incur when creating a database marketing campaign, they must review the equipment cost, security, salaries, and the many different databases involved to run hygiene and suppressions.  When you add it all up and compare it to what it costs to outsource to a specialized team that takes care of all of the overhead for you and charges pennies per name processed, then depending upon the volume and complexity of your marketing efforts, why would you not consider the results?

6. Data Sources:  Many vendors have access to a multitude of data sources.  Some are agents of one or more credit bureaus, and some also have numerous sources of demographic data available to assist with analyzing your programs or selecting marketing campaign files.  Credit bureaus frequently only provide credit information to bureau agents that they will not provide to end users.  Below is an illustration of a traditional approach to prospect mailings:

• A typical first step is to compile lists
• Run internal processing with other internal list sources
• Hygiene
• Create a mail file
• Pay for names mailed for all list sources
• Pay for matching list files and additional criteria

Traditional Approach to Prospect Mailings

With this process, you will encounter several disadvantages, including:

• Long lead times as audience selection decisions must be made early
• Data may age by the time you utilize it
• Involvement from significant internal resources to manage this process

Having a third party processor to warehouse your data enables more efficiency for each campaign through quicker analysis, as well as hygiene steps run earlier, and more accurate and faster data mining .  A third party processor will also be able to model with the data not readily available to an end user.

As opposed to the traditional approach for producing your net name file, with a third party you now have a more simplified, and efficient approach.

• The compiled list already exists
• All of the other internal processing and other list sources will be processed for you
• All Hygiene and NCOA will be completed
• Model refresh and score appends can be done expeditiously and most qualified recipients are selected.
• Mail files created
• Only pay for the names mailed, avoiding costs for multiple lists, and cleansing data
• Responses and sales can now be analyzed very quickly

Third Party Prospect Mailing

Come back later this week for the third and final installment of “What to Consider When Outsourcing Your Analytical Needs?”

By: DataLab Staff

This is a difficult question that many companies must address, and frequently important elements can be overlooked.  There are many factors for you to consider, including the cost/benefit ratio.  Frequently, organizations consider outsourcing a service that could potentially be managed internally.  However, due to various constraints; they find that their internal resources are over-taxed and struggle to adequately perform the business requirements to deliver successfully targeted marketing initiatives.

First, let’s take a look at the many advantages of outsourcing analytical needs:

1.  Timing: When outsourcing, frequently companies that specialize in database marketing services have dedicated resources that are available to make your analytical needs a priority.

2. Expertise: Outsourced organizations may have expertise that may include what your competition is doing.  Although they will not share this information, they can apply their knowledge to your benefit.  Additionally, they can also apply their experiences from other industries.

3. Database Marketing Knowledge: Corporations frequently have Analytical teams that contain a great deal of mathematical expertise; however this expertise is not always best suited for real-life database marketing applications, which requires a masterful blend of science and art.

4. Significant Cost Savings: When taking into consideration the overhead involved for internal resources, such as salaries and benefits, servers, office equipment, analytical experience, etc., taking advantage of an external resource frequently results in using less internal manpower, as you are allowing marketing savvy experts to perform the analysis, and have access to their own dedicated equipment and resources.

5. Model Retraining: A truly seasoned database marketing company will constantly be revisiting your data to make certain that the models are reflecting current marketplace trends.  They can easily set up a data transfer process with your organization to receive sales data, and can easily produce results reporting, and perform quick trend analysis that can inform you how your market may be changing, and then retrain your models accordingly.  Many organizations are too busy with their current workload to keep up with the marketplace, thus causing loss of market share.

Check back in next week to read part 2 and 3.

Deliver®

A Magazine for Marketers

Volume 7 Issue 1 / February 2011

The Model Consumers

Prospecting is making a comeback as marketers turn to new data to find the person that’s hidden beneath the persona.

by Chantal Todé

After recent years of diminished luster among some marketers, prospecting appears to be striking gold once more.

Buoyed by a host of new statistical models and other tools designed to facilitate the search for new customers, prospecting appears to be recapturing the value that it held for many CFOs before the economic downturn that began in 2007. The trend also suggests that direct marketers are once again gaining a clearer understanding of contemporary consumer behavior, after years of being buffeted about by dramatic and often incomprehensible shifts.

Yes, there was the lure of digital initiatives, too, but for many marketers, the limited success of reaching new customers was frustrating. “During 2008 and the first half of 2009, companies really pulled back on prospecting, basically because consumers stopped responding,” says Brad Rukstales, president and founder of analytic consulting firm CAC Group.

Compounding this frustration, for some, has been the realization that consumers are morphing in ways that made the traditional criteria used for pinpointing new customers increasingly less effective. “Many marketing operations haven’t been able to react very quickly to the changing consumer,” explains Aaron Davis, vice president, analytics team at DataLab, which produces large-scale direct mail campaigns for enterprises.

To continue reading the article, click here.

Make no mistake, data mining is a complicated process. It is the intricate convergence of mathematics and statistical theory, volumes of unremarkable data, intense analysis, sophisticated automation, algorithms, assumptions, attributes and, ultimately, a fair amount of good faith. Data mining is a learned art and an applied science that every direct marketer wants and needs but that only some have leveraged to its full potential.

While it is a complicated process, data mining accomplishes something quite straightforward: it unveils meaningful patterns and rules about your customers or prospects, so that you can communicate with them more effectively. If you are working with raw data — be it an internal database or data from a list vendor—understand that you are trying to use an automatic process to discover meaningful patterns and rules. Therefore, you should not impose assumptions or structure on the data, but, instead, allow it to “teach” you.

You should know to take advantage of the tools that exist for data mining. With today’s rapid increase in computer power and the increased availability of large data sets, you can unveil a lot of value within your data that you did not know exists. Traditional methods (Excel, etc.) are not powerful enough to provide intense analysis of your data, but many computing-intense analytical software programs are.

Given its complex nature, data mining is a process that can easily go wrong. With this in mind, here are a few essential ingredients to help your data mining efforts achieve success:

• Understand that 75% of the effort of a data mining project relies on the often overlooked and routine process of data preparation. Tasks such as hygiene, parsing, merge/purge, data aggregation, and the creation of derived variables must be given the proper amount of attention. This requires the largest portion of the program’s timeline but also provides the biggest payoff.

• One must understand the concept of the positive and negative outcomes. A positive outcome is a lead, a sale, and so forth, and a negative outcome is a non-response or some other unwelcome response such as a loan default. Not taking the time to properly classify the outcomes would be like taking a trip without having a destination.

• Make sure you have enough positive and negative outcomes. A good rule of thumb is to have 1,000 positive outcomes and no more than 3,000 negative outcomes. Fewer could work but cutting the positive outcomes down to 500 would increase the margin of error by upwards of 40%.

• You must supply a sufficient number of records to allow for both a developmental data set and a validation data set. You’ll need a 80% to 20% ratio of development data to validation data.

• Test and/or evaluate a minimum of three data mining software products. All programs are not created equal. Make sure you choose the software that not only fits your budget but also produces the desired outcome in a reasonable amount of time, and that can run on a system that is compatible with your operating system.

It is important to keep in mind that software is only as good as the people who use it. It is vital to make sure the person running the project has experience in data mining, which will make a world of difference in how well you succeed.

DataLab USA has had one heck of a year – new website, 1st Place Winner in the 2010 Analytics Challenge, and new clients!   Now, the Direct Marketing Association of Washington (DMAW), one of the largest regional direct marketing associations in the US, is honoring one of DataLab’s  own with the highly coveted “Emerging Leader” Award.

Join DataLab USA as we congratulate Executive Vice President, Alex Aigner, on being named one of the Direct Marketing Association of Washington’s  Emerging Leaders.

Award Ceremony will take place on Thursday, December 2nd 6PM at the National Geographic Society Building in Washington, D.C.

For more information please visit: www.dmaw.org/best-of-direct-awards

Visit Us at Booth #108

It’s the last DMA conference event of the year, but by far should be the most enjoyable (it’s Miami!).  3 days of exchanging  new ideas in analytics and database marketing with top industry experts. If you would like to rub elbows with our data experts, we  invite you to stop by Booth #108, or join us at the following sessions:

Monday, December 13th

Session: Solving the Toughest Challenges in B2B Database Marketing

Tuesday, December 14th

Session: Target Customers Effectively Through Advanced Analytics

Wednesday, December 16th

Session: Modeling for Marketers: Understanding the Value of Pragmatic Analytics

Session: Analytics Challenge: Sponsored by Hearst Corporation and the DMA Analytics Council

And if you can’t find us there … well check the golf course. See you in Miami!

Every marketer is out to get maximum postal discounts, increase ROI, and grow their database while also keeping it “clean” and up-to-date. Running National Change of Address (NCOA) every 90 days, as required by the United States Postal Service, will help you achieve these goals. However there are some marketers that think, if running a file once through NCOA is good, then running a file multiple times through NCOA must be even better. Let’s examine the relative effectiveness of single-pass NCOA versus multi-pass NCOA.

First, a quick recap on how NCOA works.  The US postal system designed NCOA ^Link to be a closed encrypted system, resembling a black box. It is designed to link individuals to their most current address, eliminating information that is irrelevant for marketers, like interim addresses. So in the example below, if we input the individual Jane Smith at an address in VA, it will provide us with Jane Smith’s address in DC, but NCOA ^Link will not provide the interim MD address.  This is done by the use of USPS record chaining.

Family Move

You do not need to run a multi-pass NCOA to achieve this result. This is a typical result from a single-pass NCOA.  However, there are 3 types of NCOA moves: individual, family, and business. It is important to know that these three different types of moves are not combined by the USPS using record chaining. All of an individual’s moves are kept separate from his/her family moves, and the same rule applies to business moves.

Going back to our previous example, let’s say Jane Smith has a sister called Mary Smith who moved with her a few times over the past few years as a family. Recently Jane got married to John Doe who lives in NY, and Jane moved there without her sister.  So the moves that are filed with the USPS are as follows:

Family & Individual Moves

If the original address on your file is Jane Smith in VA, then after running a single-pass NCOA, you’ll match to the family move records, returning the DC address. The NCOA filing for NY move was different; it was an individual move rather than a family move. In the example above, NCOA ^Link returns the current address for the family move type, which is DC not NY. In order to get the most up-to-date address, it is necessary to run through one more round of

NCOA using the output from the first NCOA. The second NCOA pass will match to the individual move returning the most current address in NY.  (See diagram below.)

The decision to run the file through additional NCOA passes depends upon the size, age, and condition of the file. The file should be large and old enough to benefit from multi-pass NCOA. For example, a 5 million record file that experiences an 8% (400,000 captured records) match rate with the first NCOA pass, has the potential to experience a 0.5% match rate on the second pass (2,000 additional moves).  If you have a small file of 300,000 records or less, then it is most likely not worth the effort and cost to run multi-pass NCOA. In many cases, running multi-pass NCOA for every marketing effort is overkill. A better alternative is to run a multi-pass NCOA on your master file every couple of years using a 48 month NCOA ^link.

What have your experiences been with multi-pass NCOA? Is your vendor recommending it? Please share comments below.