Measure two clocks, not one
StitchCapacity intentionally separates machine-run time from operator handling time. Combining them hides the bottleneck. A faster machine does not fix a hooping queue, and a second operator does not make a stitch-heavy design sew faster.
Clock A: machine-run cycle
Choose a recurring design. Record its actual digitized stitch count. Time from the moment the machine begins the design until the embroidery cycle finishes. Include the trims, color changes and ordinary pauses that happen inside that run. Exclude garment removal, hooping and loading because those belong to the operator clock.
If a 9,000-stitch design averages 13 minutes from Start to finished cycle, the observed input is about 1.44 minutes per 1,000 stitches, equivalent to roughly 692 effective stitches per minute for that design.
Clock B: handling
Time a batch of real hoop/load actions, not one especially smooth piece. Include the active work required to remove the completed item, position or hoop the next item and make the machine ready for the next cycle. Divide total handling minutes by pieces handled.
Keep flats and caps separate
Do not force one speed assumption onto every product. Caps can have different run limits, handling, registration behavior and failure modes. If caps are important, collect a separate cap sample and enter a separate cap minutes-per-1,000 value and cap handling time.
Use several representative runs
A capacity purchase should not depend on the fastest cycle you saw all week. Measure multiple normal jobs, discard only obvious abnormal events you can explain, and plan from a representative central value plus a reserve. For a small shop, five to ten ordinary cycles are already more informative than a brochure maximum.
Do not calibrate from pieces per hour alone
Pieces per hour changes with design stitch count, batch architecture and operator activity. Minutes per 1,000 stitches makes different designs comparable, while handling minutes keeps labor visible. The capacity tools can then combine those assumptions with the actual order mix.
Recalibrate after a workflow change
Measure again after adding a hooping station, changing cap frames, moving to pre-hooping, training a new operator or changing machines. The point is not to produce a universal benchmark; it is to keep your own planning inputs anchored to what the shop currently does.
What not to infer
A measured effective SPM is not a quality score and does not prove the machine can sustain that rate on every design. Dense fills, small lettering, specialty threads, caps, long color sequences and machine condition can all change actual runtime. Use the measurement as a planning input, not a guarantee.
Turn your timing notes into planner inputs
Enter your timed cycles once, add a conservative margin, then send the calibrated assumptions directly into the Capacity Planner or Queue Lab.