Where the changeover clock really starts on a boat line
The phrase changeover reduction boat building sounds like a stopwatch exercise, but on most marine lines the clock is running long before anyone pulls a part from a mold. A center console line running an 18 foot hull on Monday and a 22 foot hull on Wednesday is not just swapping tooling. It is swapping gelcoat color, catalyst ratio, cure schedule, rigging plan, and the transom cutout that decides whether one outboard or twin outboards bolt on at the end. When a plant manager asks how long a changeover took, the honest answer is usually “somewhere between the last good part off the old mold and the first good part off the new one,” and that window is almost never measured cleanly.
The problem is that boat building is low volume and high mix by nature. You are not running a stamping press making the same bracket ten thousand times. You are pulling a hull, prepping a mold, mixing a new gelcoat batch, and waiting on a gel time that shifts with booth temperature and humidity. Each of those steps has its own clock, and on paper travelers those clocks get rounded, guessed, or backfilled at the end of the shift. That is where the money quietly goes.
The hidden cost of a mold and color change
Two changeovers dominate a fiberglass boat line, and they tend to stack on top of each other. The first is the mold cycle. After a hull or deck is pulled, the mold has to be inspected, cleaned, waxed or sprayed with release, and buffed before the next layup. On a large hull mold that is often a two or three person job that runs an hour or more, and if it is done poorly the next part shows it in surface defects that force a rework nobody logged as changeover time.
The second is the gelcoat booth. Changing color means purging the gun, flushing the lines with acetone, mixing a fresh batch with the right catalyst ratio, and checking the color against a standard before the first spray. Get the catalyst percentage wrong and the gel time drifts, which either holds the mold hostage in the booth or, worse, produces a soft cure that shows up hours later. On most lines the booth is the true bottleneck, because only one part can be in it at a time and everything downstream waits on that gel window.
- Mold prep and release. Pulling the part, cleaning residue, applying and buffing release wax, and inspecting for defects, often the single longest uncounted block in a changeover.
- Gelcoat purge and color match. Flushing guns and lines with solvent, mixing a new batch, dialing the catalyst ratio, and confirming color before the mold is committed.
- Cure window drift. Booth temperature and humidity move the actual gel time away from the number on the traveler, so the “standard” changeover time rarely matches the real one.
- Rigging and transom reconfiguration. Switching between single and twin outboard layouts, wiring harnesses, and fuel system routing that changes with the model.
- Traveler reconciliation. The crew’s handwritten start and stop times get entered at end of shift, so the data you plan from is already an estimate of an estimate.
Why the numbers you plan from are usually wrong
Ask three different lead hands how long a color change takes and you will usually get three different answers, all of them confident. That is not because the crew is careless. It is because the only record is a paper traveler filled out from memory after the part is out of the booth. The gelcoat kicked at 10 something. The mold was clear by lunch. The next hull started “right after.” None of that is precise enough to schedule against, so planners build in padding, and the padding becomes the standard, and the standard hides the real variation.
The result is a line that looks fully loaded on the schedule and sits half idle in reality, or the reverse, a booth that gets double booked because two color changes were both logged as forty five minutes when one of them actually ran ninety. Until the changeover clock is measured from something other than memory, changeover reduction efforts tend to chase the loudest complaint rather than the biggest loss.
Measuring changeover from machine and system data
The change that matters is measuring the changeover from the equipment and the software instead of from the crew’s recollection. A modern gelcoat booth has temperature and humidity sensors. Catalyst pumps meter their ratio. Heated mold blankets and cure ovens run on controllers that already know when heat was applied and for how long. The mixing station knows when a batch was drawn. Each of those is a timestamp, and stitched together they draw the real changeover, from last good pull to first good spray, without anyone writing a number down.
Once the clock is measured from the line, the decisions change. You can see that color changes on dark hulls consistently run longer because the color match takes more passes, so you batch dark hulls together. You can see that a specific mold eats an extra thirty minutes of prep every time, which justifies a second release station or a mold refurbishment. You can see that the booth, not the layup crew, is the constraint, so adding layup labor buys nothing. None of that is visible on a paper traveler. All of it is visible when machine data and system data sit in one place next to the schedule.
What changeover reduction boat building looks like on the floor
In practice, changeover reduction boat building comes down to three moves that reinforce each other. First, measure the real changeover from equipment timestamps so the standard reflects what actually happens on most lines rather than a padded guess. Second, sequence the schedule around the true constraint, usually the gelcoat booth and the cure window, so molds are prepped and staged to feed it without gaps. Third, close the loop so that when the catalyst ratio or booth humidity drifts, someone knows before the part cures wrong, not four hours later at demold.
A boat plant that does this tends to find that its changeover problem was never really about how fast the crew works. It was about a booth that waits, a mold that could have been prepped in parallel, and a set of numbers that were never accurate enough to plan against. Fixing the measurement usually recovers more line time than pushing the crew ever could.
Where Harmony fits
Harmony is an AI-native operating system for American manufacturing that gets a boat plant off paper and spreadsheets and ready for AI. It connects at the PLC, Allen-Bradley and Rockwell, Siemens, Omron, Mitsubishi, over OPC UA or whatever protocol the booth controllers, catalyst pumps, and cure ovens already speak, and it pulls the mold prep travelers and rigging sheets in alongside them, so the changeover clock is measured from the line rather than from memory. That live data layer is what turns manufacturing scheduling software from a static plan into one that sequences molds around the real gelcoat booth constraint, and it is built for the low volume, high mix reality of fishing and marine production. On top of that data Harmony layers AI search, agents, scheduling, and predictive maintenance, plus back-office automation across finance, sales, procurement, and logistics. The AI proposes and a person approves, because in a plant that decision should have a human name on it. We are software and hardware agnostic, and our published pilot is $15–20K one-time over 4–6 weeks with forward-deployed engineers on-site and working software by week three. Customers include Mossberg, MoonPie, and CLS.