Industry Trends

"Looks Fine" Passes the Eyeball Test. It Doesn't Pass a Spec Sheet.

“Looks fine” is the most expensive phrase in my vocabulary

I don't trust visual inspection. I trust the spec sheet. That's an unpopular position in any company where the purchase order has already gone out and the sales team wants the shipment released. But it's the position that has saved us the most money in four years of signing off on everything from industrial adhesive batches to printed holiday catalogs.

Here's the short version of my argument: most so-called “quality problems” are actually specification problems in disguise. The glue wasn't “bad” — the spec never said how it would be evaluated. The poster wasn't “wrong” — nobody wrote down what “correct” meant beyond “Pantone 286.” Once you fix the spec, you stop arguing about taste and start arguing about measurements. That's a much better conversation.

I'll walk through why this matters, using three real categories of stuff we buy and get burned on: E6000 adhesive, printed posters and catalogs, and gift-wrap filling. Yes, those really do belong in the same conversation. Stick with me.

Argument 1: “What is E6000 glue” is the wrong question to put on a purchase order

I hear it every quarter: “what is E6000 glue, exactly?” And then someone in the room answers with a list of adjectives — industrial strength, multi-purpose, flexible, fast-drying. Congratulations. You've just written a spec that cannot be inspected.

For the record, E6000 is a single-component, solvent-based, elastomeric adhesive — it's not a two-part epoxy, despite what people searching “e6000 epoxy” are hoping to find. It skins over in around 10–15 minutes and reaches full cure somewhere between 24 and 72 hours, depending on thickness, ambient temperature, and humidity. That range is the whole point. If your spec sheet says “dries in a day,” you've given the production floor permission to interpret “a day” as loosely as they want.

Here's where I got it wrong early on. I assumed “industrial strength” was a defined term. It isn't. There's no body certifying it. I approved a batch of adhesive for our small-shipment program on that word alone, and the joints started failing in the field two weeks later. Not because the product was defective — because I hadn't specified the test method.

“I assumed ‘industrial grade’ meant a specific bond strength number. Didn't verify. Turned out it meant whatever the vendor decided it meant.”

Now every adhesive spec we issue names the substrate (steel, ABS, fabric, leather), the overlap area, the clamp time, the cure window, and the failure mode we're testing for. The word “multi-purpose” does not appear anywhere on it. If a vendor sends me a datasheet that only uses adjectives, I send it back.

What E6000 is honestly not good for

Since the whole point of a spec sheet is to say where a thing stops working — polypropylene and polyethylene are difficult without surface treatment. Silicone is a non-starter. And if you're bonding something that has to survive a hot car in August, you need to test that before you standardize it. The brand's own guidance is the best source here, and it beats whatever a supplier rep tells you at a tradeshow.

Argument 2: Your poster, your catalog, and your color tolerance are the same conversation

Now switch contexts. A boutique film-restoration client asks us to print a run of A New Hope posters — fan-art variant, but with a licensed palette. Fine. Another client wants a holiday-time catalog for a discount retail concept — think the kind of thing a Five Below shopper picks up in November. Also fine. Radically different runs, different substrates, different press setups.

Same problem: nobody wrote down what “correct color” means.

This is the one place I insist on an external anchor. Pantone's own published tolerance guidance is useful here:

“Industry standard color tolerance is Delta E < 2 for brand-critical colors. Delta E of 2–4 is noticeable to trained observers; above 4 is visible to most people. Reference: Pantone Color Matching System guidelines.”

If your spec says “match the Pantone chip,” you're going to spend a decade in revision cycles. If your spec says Delta E ≤ 2 on the Pantone 286 C and 485 C brand colors, Delta E ≤ 4 elsewhere, you now have a binary pass/fail, and neither you nor the printer has to argue about whether the blue looks a little purple today.

The poster and the catalog aren't really different problems. They're the same problem — color management and resolution — with different commercial stakes.

Resolution, briefly

For commercial offset, 300 DPI at final trim size is the industry floor. That's not a preference; it's arithmetic. The formula everyone should have memorized but doesn't:

“Print size (inches) = Pixel dimensions ÷ DPI. Example: a 3000 × 2000 pixel image at 300 DPI = 10 inches maximum width, 6.67 inches maximum height.”

I've seen a client submit a “high-res” flyer that was 1,200 × 800 pixels, stretched across a 9 × 6 inch trim. It printed. It looked terrible. Nobody in the chain could tell me whose file it originally was (circa 2021, judging by the metadata — ancient by print standards). That's a spec problem, not a talent problem.

Argument 3: Tissue paper in a gift bag is not a “small detail”

I know how this sounds. How to place tissue paper in a gift bag — really? We're discussing this as a quality issue?

We are, because when a brand ships gift packaging at scale, every unit has to unfold the same way. If forty different fulfillment staff members each have their own idea of how much tissue to use, how to fold it, and whether the logo side faces up, then the customer is receiving forty different products — and the brand is receiving forty different impressions.

We solved this with a one-page spec: two sheets of 17 gsm tissue, logo positioned top-center at 30mm below the bag rim, corners tucked behind the gift, no visible tape. I rejected the first sample run from a new fulfillment partner because — and I'm quoting the QC note — “tissue was ruffled on arrival, indicating insufficient internal support.” They thought I was joking. I wasn't. The redo cost them a weekend; it cost us nothing.

Small packaging details are where QC either earns its reputation or loses it. Nobody notices the tissue when it's right. Everybody notices when it's wrong.

The obvious pushback: “But this slows everything down”

Yes, it does. I've been told, more than once, that I'm “slowing the machine.” Fair. I'll own that. But the machine that's slowed for two days to fix a spec sheet is a cheaper machine than the one that has to explain a $22,000 redo to the CFO.

And here's the part that gets me: the data almost always says “pick the cheaper option.” Every spreadsheet I've ever run has said the same thing — the cheapest compliant quote wins. Somewhere around 2023 I stopped trusting that completely. Because the number that decides quality rarely lives on the quote — it lives in the response time. The vendor who answers a spec question in six hours and the vendor who answers it in six days are not delivering the same product, even when their spec sheets are identical.

“The numbers said go with the cheaper quote. My gut said the slower responder was telling me something. Turns out the six-day reply was a preview of the six-week delivery.”

That wasn't data. That was pattern recognition. Now I make it explicit: response time to spec questions gets written into the vendor evaluation. It's measurable. It stops being “gut” and starts being a column.

What I'd say to my earlier self

If I could redo one thing, I'd have written tighter specs earlier. Not because I knew less back then — I actually knew less, which is a terrible reason to be lenient — but because every one-off exception I approved became a de facto precedent. The spec sheet is the only thing standing between a good process and a customer telling you “well, last time you did it this way.”

So here's my opinion, restated

The best quality control document in your company isn't your inspection checklist. It's your spec sheet — and most companies don't actually have one. They have a purchase order, a reference image, and a hope.

E6000 glue doesn't fail because it's bad. It fails because “dries fast” was never given a number. Posters don't come back because the printer is incompetent. They come back because “match the color” was never given a number. Tissue paper doesn't ruin a gift bag. A process that never specified tissue paper ruins a gift bag.

I'd rather reject a delivery than accept one I can't measure. That's a bias, and I'll live with it. If you're dealing with a vendor who thinks specifying a Delta E value and a cure window is “overkill,” they're probably also going to think the redo is your problem instead of theirs. That's not a quality issue. That's the quality issue.