For a good 6 months now, my LinkedIn status has been set to: working on a combination of PlanetPress Suite, Visual Basic, and awk to automate variable data digital print. I should throw some DOS batch files in there as well – they’re part of the process that’s making it work. I guess it’s time to update my status (I should probably do that more often, eh?).
All that personalized, data-driven, image swapping to produce a one to one mailer is great. What I’m really pleased with is not the fancy pieces that go out in tomorrow’s mail, but the fact that all the data work, postal processing and template merging that went into that fancy piece of mail got done by a couple of computers while I was sleeping. Well – ok. I wasn’t sleeping – I was at my desk working on other stuff.
For me – the keys to it have been 1) Can you define your output using a template and rules, and 2) can you get consistent input data? The rest of it has simply been an exercise in stringing pieces together – with a few logic puzzles thrown in to keep it interesting.
I realized we were done with the project when we ran 13 mailings consisting of over half a million pieces combined and the most time consuming part of the process was signing the postal paperwork and cutting the POs for the postage checks. In fact, that’s all I had to do.
We took in our customer’s raw data (structured, reliable input format though) imported it into our mailing software. CASS certified the addresses. NCOA processed the addresses. Produced PDFs of the postage statements and deposited them in a client-accessible ftp folder – files named to match customer input filenames. Produced data files in postal processing order for template merging and deposited them in another client-accessible ftp folder – matching names again. Printed qualification reports, tray tags and postage statements. Sent email notifications to customer service and our client that files were done processing, here’s the record count and here’s where to find the postage statements. For this particular client, they pick up the data and do some further work on it before we print, but the data could just have easily been template merged at that point and sent down to one of our digital presses. Hands off from raw data to files queued up on the RIP ready for printing.
PlanetPress has some excellent flexibility for designing templates, so I’m thinking the next weakness to address is reliable data input. Maybe the next step is making input buckets for various input formats (FullName in a field vs FirstName, LastName in separate fields). On the other hand, talking with an acquaintance who writes log analysis software – maybe the answer is tokens. Similar to describing your Apache log file format in httpd.conf. Give me a token string that describes your data, and I’ll know how to parse your data correctly. Hmmm – check back in a few months – I’ll keep you posted.