The problem is not the press
I’m a quality/compliance manager at a print-on-demand company. I review around 250 unique jobs every month—books, catalogs, brochures, posters, envelopes, and the occasional odd project. Four years of doing this has pushed me toward a mildly unpopular conclusion: most quality problems are not printing problems. They are specification problems.
Presses can drift. Cutters can go dull. Binders can jam. But the failures that cost real money happen when someone described a job in words instead of measurements. “Poster” is not a specification. “Nice thick paper” is not a specification. “Make it look good” is not a specification. If you don’t define good, quality control has no basis to say something isn’t good.
Clear specs are not bureaucracy. Clear specs are efficiency.
I don’t have hard data for how many first-pass rejects across the industry start with vague copy rather than mechanical faults. Based on my own order queue, though, my sense is it’s easily more than half. And that part is fixable.
The phrase tax is real
A while back, we got a project request that said, simply, “the summer hikaru died poster.” The file attached was gorgeous. The specification line underneath was not: it said “poster of attached art.” No dimensions, no paper, no finish. By the time we clarified whether the client wanted a 12×18 locker poster or a 24×36 collectible print, the job had lost two business days—and any chance of a quick turnaround. That delay didn’t happen because the press was slow. It happened because the phrase tax was due.
This is why I don’t roll my eyes when someone asks about the “Lightning Source login.” If you’re an independent author or publisher, the portal is genuinely useful. But the login is not where quality happens. Quality happens when you fill in the fields after login: exact trim size, page count, paper stock, ink type, finish, bleed. Those fields feel like red tape, but they’re the shortest path to a product that looks the way you see it in your head.
We work inside the same print-on-demand ecosystem as Lightning Source/Ingram, so I’ve watched files travel from a bookstore order to a production queue. When the metadata matches the PDF, everything flows. When it doesn’t, the automated system doesn’t magically fix it. It spits out a proof that somebody has to question, revise, and re-approve. That’s the opposite of efficiency.
Search terms are not specifications
The same vagueness shows up outside printing. People search for “window film I can see out but not in” and expect a single answer. But a one-way privacy film depends on light levels, visible light transmission, film thickness, and reflectivity. If you don’t specify a percentage, you’re not choosing a material—you’re hoping for one. And a phrase like “how thick is electrical tape” is not a spec either. General-purpose tape is often around 3 mil; heavy-duty professional tape can be 7 mil or thicker. The difference affects dielectric strength, abrasion resistance, and where you can safely use it. Without a number, you’re comparing guesses.
Printing behaves the same way. “I want a book” could mean a 5×8 perfect-bound paperback or an 8.5×11 coil-bound manual. “I want it cheap” doesn’t tell a vendor whether a 60 lb paper will work or whether you need an 80 lb opaque to stop show-through. If my team isn’t given measurable specs, we either guess or send the job back. Both eat time.
I still hear “this feels like overkill”
I get the pushback. Local short-run printers exist for a reason, and some of the best custom results I’ve seen came from a person walking into a shop and talking through options with a production expert. I’m not asking anyone to replace that with a 14-page specification document.
But here’s the difference: a local conversation can absorb ambiguity. A global print-on-demand network can’t. If a customer orders a book through a bookstore website, that title might be printed in Tennessee, or in the UK, or somewhere else in Ingram’s network. The file has to carry its own instructions. If you search for “Lightning Source Sharjah” because you’re trying to find a nearby plant, I understand the instinct—shipping costs matter. But location doesn’t solve an unclear spec. I can’t verify every facility in every distribution network, and I don’t need to in order to say this: a portable, well-specified file is better than any local workaround.
There is also an automation benefit. When specs are explicit, we can set pass/fail tolerances and let the quality team focus on the handful of jobs that need judgment. That’s how we shortened proof approval without raising defect rates. In Q1 2025, our first-pass yield improved by 4% after we started asking clients to submit a simple spec form. The format felt overly strict to a few authors, but the results were hard to argue with.
Start with the spec, not the hope
Here’s a surprisingly useful example: USPS defines a standard letter envelope as roughly 3.5 × 5 inches minimum up to 6.125 × 11.5 inches maximum, with a thickness no greater than 0.25 inch. If you design an envelope that falls outside those dimensions, you’re not mailing a letter anymore—you’re mailing a flat or a parcel, and the postage changes. According to USPS Business Mail 101, the physical dimensions determine the price. I like this example because specs aren’t a creative limitation; they’re a language of consequences.
The same principle shows up in the cheaper-spec trap. I had a client pick a lighter uncoated paper for a 300-page novel to save about $0.30 a book. The first physical proof looked like text shadowed through the page from the other side. The show-through was so distracting that they ended up spending $2.10 more per book on an 80 lb opaque stock, plus rush charges, to replace the copies that had already been printed. That “low cost” paper was one of the most expensive choices they made all year.
If I could redo any decision on that project, I would have asked a basic question before ordering: what opacity do we need for this page count? But when budget pressure kicks in, it’s easier to hope a thinner stock will be fine. It wasn’t. (Mental note: I really should build opacity checks into every project template.)
So here’s where I land. The most important quality control step happens before a file ever reaches the press. It happens when you write the spec. You don’t need a graphic design degree to do it—you just need to answer the questions: size, stock, finish, quantity, color. If you can’t answer them yet, ask someone before production, not after.
Next time you upload a file, whether it’s a book through Lightning Source or a poster based on moody anime artwork, give the system something to measure. Don’t just describe the product you want. Specify it. In my world, that’s the difference between hoping for quality and actually inspecting it.
