by Patrick Helskens
When it comes to variable data printing, I am confident that everybody in the printing industry has some understanding of the term. However, when analyzing customers’ applications and conversations on the matter, I feel that there are still a lot of misconceptions. Not only on what variable data actually is but also on the complexity to implement it. Let alone, the benefits that come with it when implemented correctly. Let’s settle this once and for all, shall we?
Let’s start by defining what variable data printing is. Here is what I found on Wikipedia:
Variable data printing (VDP) — also known as variable information printing (VIP) or variable imaging (VI) — is a form of digital printing, including on-demand printing, in which elements such as text, graphics and images may be changed from one printed piece to the next, without stopping or slowing down the printing process and using information from a database or external file.
An example in this definition would be a set of personalized letters, each with the same basic layout, printed with a different name and address on each letter.
Variable data printing is indeed mainly used for direct marketing, customer relationship management, advertising, invoicing and applying addressing on self-mailers, brochures, or postcard campaigns.
So, this Wikipedia assessment is a decent definition indeed.
But does it really cover the load? Let’s dig a little deeper and go into the specifics.
Variable. That’s obvious, the print content will vary based on a set of rules.
Digital printing. Very true. Without going into detail on the exact equipment used, changing print content variably can only be accomplished by using a digital printing press (or inkjet print heads mounted on some other device).
Elements. A tentative yes. There is often more to it than just text, graphics, and images. (More about that later.)
Database. Absolutely. When creating variable content databases are bound to come into play. Not always though. Other ways exist — sometimes much simpler — to drive variability. (Again, more on that later)
Personalized. Okay. This is when you print variable information that is only correct or relevant for a specific “person”, right?
This for sure defines personalization as we all know it. Nevertheless, Variable Data Printing is sometimes inadvertently perceived as a synonym for personalization. And that is not true. Personalization is a mere subcategory of variable data. In my book, variable data is a much broader concept, going beyond adding the name and address and unique customer ID of a person onto a fixed background.
Main use. I object. Direct marketing, advertising, invoicing, addressing, are not the only “main” applications of variable data printing. They are undoubtedly the group of applications most easily associated with the variable data concept. But limiting the scope of what can be realized by printing digitally, does not do the concept of VDP justice.
What lies beyond personalization? Variable Data and customization!
Variable data printing also enables the customization of documents. This could mean sending thousands of tailored (entirely or partly unique) messages to just as many individuals. This type of personalization wants to maximize a target’s interest and response. Such direct mailing and transpromo campaigns typically combine a variety of variable fields, such as:
The functional personal elements: name, address, ID, postal bar code, date.
The general message: typically, a pool of pre-defined text blocks (and sometimes backgrounds) that don’t target a specific person, but rather a group of people belonging to the same category. This is what we call versioning.
A call to action: a unique QR code to register onto a website, a personalized product offering, a temporary promo code, or the overview of spending behavior linked to a specific discount, or variable product coupons with changing values.
Contact details: the picture of the branch closest to your location, the picture, and details of/for your local contact person.
"Variable Data Printing is sometimes inadvertently perceived as a synonym for personalization. And that is not true. Personalization is a mere subcategory of variable data. In my book, variable data is a much broader concept, going beyond adding the name and address and unique customer ID of a person onto a fixed background." – Patrick Helskens
All of these elements are visible to the receiver. In today’s world they instinctively “feel” that this document was specifically made for them. But until they get to compare their document to someone else’s, they remain in the dark as to if and why the piece was customized.
On the sender’s side, it is much easier to measure interaction, to further tweak messages, and prove the augmented response of such a customized approach compared to a static mass mailing campaign.
And there is still more than meets the eye: there are numerous elements that are required during or after the production process that are needed to finalize the process but are often invisible to the receiver. After printing customized documents, the print provider still needs to cut, count, collate, fold, sort, and put them into envelopes or convert them into a finished product. The output may range from complete book blocks or sets of lottery tickets, or wallpaper murals or Point-of-Sale price tags bundled per store… The possibilities are endless.
This final part is the industrial side of variable data printing and is equally important.
For example, based on my personal buying behavior, my envelope will contain two folded sheets, while someone else will get five. Based on my profile I will receive an eight-page newspaper without the sports section but including science articles, while my colleague prefers stock market information and gets a 12-pager.
So, it’s not only the content of the document that is variable per person; it can also be the number of sheets. Most finishing devices are equipped with simple mark readers or bar code scanners to identify what is passing by. The digital press will add these based on certain rules. Several data matrix codes and other markers (variable of course) are used to count all the sheets, identify sheets within a group, start/stop collators, and drive enveloping machines.
Also, human-readable alphanumerical data is used to uniquely identify sheets in case of issues. The more complex and variable a print run becomes, the more important it is to have very accurate track & trace capabilities.
Truth time - how hard is it to make a variable data job?
It depends on the method you choose. Here are your options:
Load a static document in the printer’s memory and instruct it to print variable data on top of each static document. This is the simplest way to produce VDP, however the capabilities are limited.
Another method is to merge static and variable elements using standard pre-press software. These authoring tools can then export the complete job to a static multi-page PDF. This produces potentially huge files, with all elements being merged into every page.
The downside of this approach is that someone needs to spend quite some time in pre-press to create and generate the job into several batches. Running many such very large print files can easily overwhelm the print system’s processing capability. When this happens, it becomes impractical for the RIP (Raster Image Processor) to deal with print jobs of more than a few hundred pages.
Let’s say that a RIP can process one duplex A4 per second and that the print job contains 5.000 A4 documents, then that results in a RIP time of 5.000 seconds. That is a wait of almost one hour and 40 minutes before printing can start!
A third method is to combine the static and variable elements into output files, prior to printing, using specialized VDP software. This produces more optimized files, such as PDF/VT, PostScript or PPML, which maximizes the RIP’s capability because it only needs to process the repeating static and variable elements once.
However, also here, someone in pre-press needs to design the pages, export them into one multi-page file and then RIP it all on the printer/press before the production can start.
And then there is the fourth method. Specialized high-end “programming” tools that generate IPDS (Intelligent Printer Data Stream) or AFP (Advanced Function Presentation) output offer a way to get around RIP processing time limitations for very complex variable data jobs. But since this is such a specific, expensive, and centralized setup for producing huge variable data print runs, we can skip this for now.
It is not diﬃcult to create a variable data application. But as always, it starts with a good description of the requirements and expectations. Successful (variable) digital print applications are those that are specifically designed for digital; intelligently combining the correct pre-press authoring tools, the databases (or MIS/ERP connectivity in combination with JDF/JMF), the press’ digital front-end capabilities and specific print features, and the in-line or near-line processing afterwards.
Printing variable data applications on a digital press is not more expensive (for the print provider) than printing a static job. However, there is an extra cost for the customer, which stems from the design stage and maybe the track & trace requirements added later on.
Why is the Xeikon workflow your best bet for Variable Print Jobs?
The Xeikon workflow helps generate variable data jobs without taking a lot of time or extra effort. This workflow does not require pre-press tools and time, and it optimizes precious RIPing time by automating rules “on-press” that normally would have to be done up-front.
The Xeikon X-800 workflow, running on a standard multi-core Windows platform, is object-based. That means variable backgrounds, variable images, variable text and bar codes, finishing marks, separator sheets, logistical data – all these, potentially repeating, elements are processed only once and re-used (referenced) multiple times, or they are added on the fly in the workflow itself without being part of the input files.
Once RIPed, these become (RPEs) ready-to-print elements to be “picked” based on a set of rules. No time is lost on imposition lay-out, adding the screening, applying tone curves, changing orientation or print sequence, defining start-stop pages, injecting separator sheets, etc. All that is done after the RIP process, and it is virtually instantaneous.
Multiple RIP instances can work together, either to distribute the pages in a large multi-page file over several RIPs, or to process multiple jobs simultaneously, or both.
In short, if a customer can describe the required behavior of all the variable elements, rest assured it can be automated.
This process does not mean that specialized pre-press tools are obsolete. The X-800 workflow is used for automating variable data, not for complex graphical design. Placing a variable text on a wavy line is a design function, not an automation function.
The attentive reader realizes we are coming full circle.
Time to put our money where our mouth is. Let’s illustrate with some examples to show you the brilliance of the XEIKON X-800 workflow.
A high-end security printing example of a tax stamp application
Input pdf: One 42 x 20mm (landscape) fixed background containing security designs (guilloches, logos, text) defined in 5 named color planes.
Automated actions in X-800 (no operator intervention):
Preflight the pdf background for errors? Less than 1 second.
RIP background once (native 1200dpi LW) + re-map the 5 color planes to press installed spot/process colors + virtually “duplicate” background 80.000 times? Six seconds.
Add variable data from two separate database files; 80.000 OCR A alphanumeric codes & 80.000 alphanumeric records converted on-the-fly into a QR code = 160.000 variables? Two minutes 36 seconds.
Imposition: rotate 90° (portrait) + order the 80.000 unique tax stamps into 20 full continuous lanes (seamless roll-to-roll printing) + add a 0,5mm wide black continuous stripe left of the imposed artwork for finishing registration purposes? 19 seconds.
Extra automatic additions after the imposition: a 2mm wide continuous stripe on the left side of the web for finishing registration purposes + a horizontal registration mark every 10 rows of stamps + 2 vertical registration marks centered on every 10 rows length (left & right) + before the job, a start separator sheet followed by 5 meters of blank substrate (containing registration marks) + right after the job, 5 meters of blank substrate (containing registration marks) and a sheet identifying the content of each lane of tax stamps (a QR code concatenate of the first and last stamp serial in that lane) and a stop separator sheet? 25 seconds.
Conclusion: 80.000 unique tax stamps containing 160.000 variables and a bunch of technical stuff around it based on a set of production rules, as requested by the customer, without pre-press tools, starting from 1 (one!) on-press background pdf, hands-off. Ready to print in 3 minutes and 26 seconds.
And once the setup for this job was clearly defined, then the next run simply requires updates of the linked databases and the corresponding copy count.
Feeling lucky? Let’s have a look at this lottery ticket application
Input pdf: 17 different fixed lottery ticket files of 47,7 x 98,5mm (Big winner 500$.pdf, Big winner 250$.pdf, Small winner 50$.pdf, Small winner 20$.pdf, Small winner 10$.pdf, Small winner 5$.pdf, Small winner 1$.pdf and 10 different Loser designs 0$)
Automated actions in X-800 (no operator intervention):
Preflight all 17 pdf files for errors? Less than 1 second.
RIP all 17 tickets once and duplicate each ticket winner type multiple times based on a database file (post-RIP) - e.g. 2x Big winner 500$, 2x Big winner 250$, 4x Small winner 50$, 10x Small winner 10$, etc… Total time for one complete lottery game consists of 3.990 tickets? 16 seconds.
Add variable data in the bleed area of all 3.990 tickets, following this rule: [ticketname.pdf ticketcopycount/tickettotalcopies] = 3x 3.990 = 11.970 variables? 3 seconds.
Imposition: layout all 3.990 tickets on sheets of 508 x 711mm using a special template — “cut & stacked vertical shuffle: 10 horizontal & 7 vertical” — that will randomly shuffle all the tickets over a total of 57 signatures? 6 seconds.
Extra automatic additions after the imposition? Do a vertical density strip on each press sheet left of the job and add a cutter mark with an identification datamatrix alongside a sync mark for further finishing. 3 seconds.
Conclusion: one lottery game of 3.990 randomized tickets containing 11.970 variables and a bunch of technical stuff around it based on a set of production rules, as requested by the customer, without pre-press tools, starting from 17 input pdfs, hands-off. Ready to print in 28 seconds.
Each time the job is re-processed the tickets will be re-shuffled and randomized over the 57 press sheets, meaning that each game stack is different.
Child’s play? It just might be for this children’s book application.
Input pdf: 1 file containing 36 pages.
Automated actions in X-800 (no operator intervention):
Preflight the multi-page pdf for errors? Less than 1 second.
RIP the book file once? 6 seconds.
After RIPing 1 book, load a database file of 10.000 lines, duplicate the book 10.000 times (= 360.000 virtual pages) and add 4 text lines to the first page of each book, following these rules:
Line 1: “Hello [child’s name]!”
Line 2: “This is your personal copy. Enjoy.” (fixed text line for every child)
Line 3: “You have [child’s name frequency] name.”
Line 4: “There are [child’s name counter] other [child’s gender] called [child’s name] in your county.”
This is your personal copy. Enjoy.
You have an extremely unique name.
There are 6 other boys called Fitzroy in your county.
The result: 10.000 books of 36 pages, containing 50.000 variable database fields and 10.000 fixed text fields. All in 1 minute 45 seconds.
Imposition: layout the 10.000 unique books (360.000 pages) using a 2-up saddle-stitched template, generating 90.000 duplex signatures. 1 minute 17 seconds.
Extra automatic additions after the imposition? Add a cutter mark with an identification datamatrix on all 90.000 press sheets with start & stop datamatrix codes to identify every first and every last sheet of a book for further finishing. 1 minute 11 seconds.
Conclusion: 10.000 personalized books, 50.000 variables, 10.000 fixed text lines, start/stop codes, and some technical stuff, as requested by the customer, without pre-press tools, starting with 1 input pdf, hands-off. Ready to print in 4 minutes and 30 seconds.
And once the setup for this job was clearly defined, the next run simply requires an update of the linked database and the corresponding copy count.
What do you think? Sound too good to be true?
I assure you it’s not.
Call my bluff, let us demo it for you. It’s easier to show this process than to write about it. Come to one of the upcoming Xeikon Café sessions.
Did I leave a stone unturned when it comes to VDP? Let me know.