Bit by Bit: The Ins and Outs of Variable-Data Printing for Designers

1

In last week’s column, I suggested that variable-data printing is a gold mine for creative professionals willing to learn the special combination of skills necessary. Variable-data printing, if you’ll recall, uses design tools, database technology, and digital presses to create custom pieces that target specific users according to a set of criteria. In my previous column I cited the example of a car dealer mailing special offers for Volkswagen repairs only to owners of certain models within a certain region. This approach represents a more efficient way to attract the right customers than just sending mailers willy nilly to all car owners in the area.

In this follow-up column, I’ll describe some of the software that’s available and fill you in on a few “gotchas” that designers new to variable-data printing should know about when working in this arena.

Various Applications
The software used to produce pieces destined for this form of printing build on your existing skills. Creo Darwin and Datalogics DL Formatter are XTensions to the popular QuarkXPress page-layout application. Datalogics also sells an InDesign version. Atlas Software PrintShop Mail is a stand-alone application that works with templates created in EPS or PDF format from any application. (Functional demo versions of these applications — which cost upwards of $3,500 each, depending on usage — can be downloaded from their respective Web sites.)

These applications merge the data into the page-layout to create one-of-a-kind products ready to print. In some cases you will print directly on creation, in others (large files destined for the specialized digital presses) you create a massive collection of files that is delivered to your printer-partner for output. Last month I created a 21,000-page PDF file as a product of variable-image file prep! (Even at roughly one record per second, it took more than five hours to prepare this document for print. I could not use the computer for anything else during that time because it slowed the process.)

File preparation is much more time consuming that an equivalent static job. Numerous tests must be carried out to ensure that the data merges correctly with the variable-data software, and learning to mix the commands necessary for variable-data production with the preparation of a standard document takes time. My experience with DL Formatter, for example, required days of study just to get the concepts and learn the interface. With iterative improvements I was eventually able to make some good variable-data work come out the other end.

One mistake that developers of this software make is to think too small. I fear that they have a “small production” mindset, which causes large production projects to clog the arteries of production and fail. My huge database documents were too much for several systems tested. Printing small numbers locally was much easier, and more in line with the capabilities of the software.

Luxuries and Limitations
In my experience, Creo Darwin is the darling of variable-data work. It has a charming interface, excellent control over variables, and a several other features that make it shine. It can, for example, generate serial numbers, or random numbers on the fly. The other applications I tried will only create this type of work if your database generates the serial or random data prior to production. DL Formatter is very robust yet quite quirky. PrintShop Mail is easy to use, but because it relies on prepared templates, slightly more limiting.

All of these applications foil any attempt at “proper” typography in variable fields. If you like kerning — forget it! All in all, they each have the power to produce good-looking variable printing, and produce it consistently if you are willing to overlook aesthetics in favor of getting the job done.

I would like to see all of these applications advance to the next level of efficiency, and allow me to open and use multiple databases simultaneously within the variable-printing environment. I would also like to be able to apply more sophisticated Boolean functions in production. DL Formatter, for example, can only perform one set of “if-then” functions. I would like to see it support “if-then-else-and-or-not functionality.” To get it to work to my more sophisticated requirements, I had to pre-process the data to make the Boolean decisions, then export a flat database file with this data already complete. In some cases I had to create new fields in my database, and then create moderately complex calculated and relational fields to get the results I needed. There is ample room for improvement for this kind of software.

Now let’s explore territory that may be unfamiliar to new users of variable-data printing, such as mailing-list acquisition, post-office requirements, and image-substitution strategies.

Unvarying Good Design
Design for variable-data printing is the same as any other good design, but the variable elements must be carefully planned and constructed. To prevent your mailing from looking like the hackneyed work of the Publishers Clearing House, pay attention to good typography and quality images, and use your variable-data-printing software to change more than just the name and address. Changing photos, backgrounds and graphics can make a mailer much more effective than dorky “boilerplate” text changes alone (I received one recently addressed, simply, to “Dear ,”).

If you are marketing to upscale BMW owners, use upscale images and classy design. If you are attempting to get a person who plays golf to take a vacation in Hawaii, then consider featuring a photo of a golfer in Hawaii — and make sure that you use a woman playing golf when mailing to women, and vice versa.

Don’t Call Me Comma!
Don’t insult your recipients! When my name is mangled, or not included, I just throw it out. Good mailing lists can be generated that “populate” data fields so that residents of San Luis Obispo (or San Juan Capistrano or San Juan Bautista), California are not truncated into unintelligible monosyllabic grunts. Note to mailers: I don’t live in SN LUS OBSP! (The full name of our town is San Luis Obispo de Tolosa, a nightmare for database people.) This truncation is a legacy of the age of punch-cards, and there is no reason that modern-day databases must be scrunched like this.

Failure to get the names right undermines your efforts.

More than Variable Text
Successful variable-data mailings require more than just text substitutions — you need to think about pictures, too. Variable-image printing requires not only a good database, but also a library of appropriate photos that can be substituted in flight according to the demographics of your recipients.

In one recent project, I amassed 26 different automobile photos that I categorized according to their appeal to different age groups. I put Cadillacs in my “older” age category, and Audi TTs and Porsches in the “younger” age category (though “older” people can more easily afford the “younger” group’s cars). I ended up with a complex system that divided my recipient group of about 10,000 people into age, and gender-specific groups. I then pared that list into regional groups based on ZIP codes (assuming that inlanders are less interested in convertibles, and that four-wheel drive vehicles get real “traction” in mountainous regions). In the final output I was pleased to see variability in nearly every copy according to my complex matrix of recipient characteristics.

Preparing photos for variable-image printing is best done if every replaceable image has exactly the same resolution and proportions as its counterparts. This facilitates the process of printing, and saves time when producing your final project. Triple-check your images to ensure that they are all in the same color space (I found that CMYK TIFF images cause less trouble on most systems). This will prevent slow-downs and failures in production.

The Mailing-List Jungle
Acquiring good mailing lists is a real challenge. Ask anyone who has attempted to do this kind of printing, and they will tell you that the database was their biggest hassle. Good mailing lists are available for almost any group, but are usually “dirty,” meaning that data elements are found in the wrong fields in the database, or many duplicate elements exist. A clever database practitioner can import, merge, filter and produce clean, effective data for variable-data printing projects, but it is not a simple task.

Unless you’re a closet database geek, it’s best to partner with a seasoned pro on this.

ZIP It into the Mail
The Postal Service will discount postage if your mail is ZIP-sorted. They offer deeper discounts if you do detailed sorting. USPS consultants who can help you execute this sorting process are available in regional offices. Adding the PostNet code to the bottom of the addressee panel also results in a postage reduction (PostNet is the little picket fence below the address on most mail as shown in Figure 1. It has the recipient’s ZIP+4 code and a “checksum” digit in it).

I managed to generate compliant PostNet coding with a special type font in my projects, and this makes the Postal Service happier. My local BMW dealer who sent me an offer erroneously could have reduced its postage costs significantly just by ZIP-sorting its envelopes and working with the USPS for a better rate.

Opportunities Abound — Variably!
Examples of the possibilities for variable-data printing include flyers and promotional media (credit card companies could do a much better job, in my opinion), sales of upgrades to existing products (you already know who the clients are), tickets with sequential numbering, security badges with randomly-generated bar-code elements, travel sales literature based on travel records of existing customers, and much more.

Curiously, many of the examples of variable-data printing we see are based on banking examples. A bank client of mine told me recently that this is ironic because banks are prohibited by Federal banking laws from using any of the financial information they have about their customers. Our banking information is private, and that’s good. The BMW dealer, however, is free to make targeted-marketing products that appeal to their customers and potential customers.

I think that variable-data printing is a vast untapped market, waiting for the right people to create businesses that can combine the talents of graphic designers with the skills of good marketing, and the work of a clever database manager to deliver targeted marketing pieces that are inoffensive yet spectacularly effective.