The Four-Vector Screen: Where the Money Hides in a Custom Injection Molder

Reading time: 11 minutes. What you'll see: the screening method we run before proposing anything, applied end to end. Not one problem priced, but four vectors ranked against each other, because the most expensive mistake in operational AI is building the second most valuable thing first.

The company

Osprey Molded Products runs twelve presses in Columbus, Indiana. Forty-five employees, roughly $11M in revenue, custom injection molding with light assembly for appliance, industrial-controls, and outdoor-equipment OEMs. Owner-led, profitable, modernized: a state matching grant helped fund two new presses and a materials handling upgrade three years ago. The floor is not the problem. The question the owner cannot answer is where the next dollar of improvement actually lives, because four different problems all feel urgent and nobody has priced them against each other.

That pricing exercise is the screen. Two weeks, four vectors, every number labeled, one ranked answer.

Vector one: the scrap gap

Osprey quotes 4% scrap. Published benchmarks say measure before believing: industry data puts average plastics scrap near 5% with measured totals, including startup and changeover purge, often running closer to 10%, and processors have been reported losing 4 to 7% of revenue to scrap. We count everything for two weeks: inspection rejects by defect type, purge parts, regrind that never returns.

Vector two: changeover

Custom molding means changeovers, and changeovers mean purge and settling time. Published breakdowns count 25,000 to 40,000 startup-and-purge parts per press per year in operations of this shape. Osprey runs about nine changeovers a week across the twelve presses, averaging 3.1 hours each, measured, not the 2.0 hours the schedule assumes. The gap is 1.1 hours × 9 × 50 weeks ≈ 495 press-hours a year, worth roughly $37,000 at Osprey's loaded press rate ($75/hour, assumption, labeled), before the purge material already counted above.

Vector three: the demand signal nobody reads

Osprey's three largest customers are seasonal, and the plant discovers demand swings when purchase orders arrive, then scrambles: overtime in the ramp, idle presses and built-ahead inventory in the trough. The customers' own release schedules and portal data contain the earlier signal; nobody has time to read three portals weekly. We model the exposure conservatively as the carrying cost of inventory built early plus ramp overtime, $28,000 to $61,000 a year (assumptions labeled in the full model), and we flag this vector honestly: it has the widest error bars of the four, because it depends on customer behavior we can only sample.

Vector four: the two heads that hold the plant

Process setup for Osprey's 300-plus active molds lives in two people: a process engineer and a senior setup tech. Their settings sheets are years out of date; the real knowledge is theirs. This is the vector owners feel and never price, so we price it the only honest way, as risk: the measured cost of one messy departure (recruiting, ramp time at published wage benchmarks, and the scrap spike every molder who has lived it will recognize) ranges $60,000 to $140,000 as a one-time exposure, probability not modeled, mitigation cheap relative to the tail.

The ranked answer, including the row where this fails

The screen's product is the ranking, not any single number. Osprey's owner walked in believing changeover was the fire; the count said scrap carries four times the money, and the two share a build. That reversal, bought for two weeks of measurement, is the entire point.

What this costs

The screen above is our $2,500 audit, fee credited toward the first build within 90 days. The scrap-and-changeover build lands in our automation tier, from $7,500, live inside 75 days or the money back. Against the ranked table, payback lands inside the first year on the scrap vector alone, and the method's counting stage is published in our worked audit example, with the quoted-versus-measured scrap evidence in our plastics scrap analysis.

Want the four vectors screened on your plant? Start with the free 30-minute scan, or read how we measure first.