When orders begin to outrun the week, the first proposed fix is often the most visible one: a larger machine, a deeper supply order, or another person on the floor. Any of those could help. Any could also move cash into the wrong place while the real delay stays untouched.
Useful production capacity planning starts with a harder question: what evidence would prove each option wrong? Before committing to equipment, inventory, or labor, run three small tests designed to isolate the actual constraint.
Build one eight-shift capacity baseline
Choose one important product family and follow it through eight representative production shifts. Record what becomes acceptable finished output, not merely what gets started.
For each shift, capture:
- demand due during the period;
- units ready at the start;
- units released at the end;
- time spent producing, setting up, cleaning, waiting, reworking, and checking;
- the largest queue and where it formed;
- any material or packaging shortage;
- any task waiting for a person, approval, or machine; and
- the reason finished units could not be released.
Use one reason for every meaningful delay: equipment limited, material starved, labor waiting, quality hold, setup, priority change, or other explained exception. Eight shifts are a bounded starting point, not a universal sample size.
A real manufacturing bottleneck should appear repeatedly. One late delivery, long setup, or late founder night is not enough to prove which investment belongs first.
No-buy test 1: challenge the equipment case
Equipment deserves attention when a stable step repeatedly runs near its usable time and good work queues in front of it. Test the claim without buying the full asset.
Time a fixture or layout change. Rent or demo a machine with the real product. Use a contract service for one controlled run. Compare good units per hour before and after setup, cleaning, changeover, inspection, and ordinary interruptions.
The equipment case weakens when the machine is idle, material kits are incomplete, approvals interrupt work, or downstream labeling and quality checks cannot absorb more output. Rated speed is not added capacity if extra units cannot be released.
Record verified additional good units, surrounding capacity, installation and training needs, maintenance, and the next likely queue.
No-buy test 2: challenge the inventory case
An inventory investment can protect capacity when the same approved ingredient, component, bottle, cap, label, or carton repeatedly starves the constrained step. More stock is not a cure for inaccurate counts, uncertain demand, or materials that are available but not released for use.
For one high-impact item, test a temporary reorder change or create complete production kits before the shift. Track whether the added availability prevents stoppages and increases released units. Include supplier lead time, order minimums, shelf life, storage, inspection time, common use across products, and cash tied up.
The inventory case weakens when shortages change from item to item, the supposedly missing material was actually on hand, excess finished goods are already waiting to sell, or production remains blocked after the item arrives. In those cases, better records, purchasing discipline, priorities, or release rules may matter more than a larger order.
No-buy test 3: challenge the labor case
Labor is a strong candidate when ready work waits for trainable hands or a scarce skill while equipment and approved materials are available. Before a permanent hire, isolate a documented task for cross-training, temporary help, overtime, or a short scheduled shift.
If the debate is whether to buy equipment or hire, make both tests face the same outcome: additional acceptable units released per week. Do not compare machine nameplate speed with employee hours. Compare the whole flow after training, supervision, breaks, setup, quality review, and rework.
The labor case weakens when the new person spends most of the test waiting for instructions, materials, workspace, or founder approval. That result may point to a production-handoff problem instead. Use the production handoff readiness guide before turning undocumented founder knowledge into a job description.
Compare the three tests on evidence, cash, and reversibility
Put the results on one page.
| Decision question | Equipment test | Inventory test | Labor test |
|---|---|---|---|
| Which recurring wait should it remove? | Physical processing limit | Repeated material starvation | Ready work waiting for hands or skill |
| What did the small test change? | Good units per usable hour | Stoppage hours and kit readiness | Good units released per trained hour |
| What cash becomes fixed? | Purchase, installation, tooling, maintenance | Stock, freight, storage, expiry risk | Recruiting, wages, payroll burden, supervision |
| How reversible is the next step? | Often low after purchase | Depends on common use and shelf life | Higher with temporary or limited-hour coverage |
| What becomes the next constraint? | Finishing, cleaning, QA, space, or power | Storage, inspection, or cash | Training, workspace, or approvals |
The smallest test should show whether the proposed investment changes the measured queue and increases acceptable finished output without creating an unacceptable quality, cash, or safety problem.
Worked example: three ideas, one confirmed constraint
Consider a hypothetical product business with repeated demand for 480 units per week and current released output of 360. Across eight shifts, mixing can support 600 units, packing can support 520, and approved materials are available. Filling is limited to 15 good units per hour across 24 usable hours: 360 units. Work repeatedly queues before filling.
The business runs three bounded tests:
- staging extra packaging changes no stoppage time because packaging was not starving the line;
- four hours of packing help moves completed units faster but does not increase filled output; and
- a rented filling fixture raises the tested rate from 15 to 20 good units per hour.
At 20 units per hour across the same 24 hours, filling capacity becomes 480 units. The useful gain is 120 units because demonstrated demand is 480, even though packing could handle 520. The equipment test has not proven that a purchase is affordable, safe, or wise. It has shown that the equipment option directly affects the confirmed constraint while the other two tests do not.
The next review should compare full cash requirements, training, maintenance, quality checks, downtime risk, and whether packing becomes the next queue. The healthy-growth cash and capacity bridge can help keep higher output from becoming the only decision measure.
Keep daily priorities separate from capacity investment
A capacity decision solves a recurring structural limit. A dispatch decision decides what ready work should run today. Use the readiness-first production dispatch board when several jobs compete for the same shift, but do not mistake a priority conflict for proof that another machine, larger order, or new role is required.
Frequently asked questions
How long should a capacity test run?
Use enough representative shifts to observe the same limit more than once. Eight shifts are a practical starting point, but seasonal demand, frequent product changes, or rare supplier delays may require a longer comparison.
What if two constraints appear together?
Run the least fixed test on the constraint closest to finished output or the one that most often stops the whole flow. Re-measure after the change; the second constraint may become the next priority.
Can inventory be the right first capacity investment?
Yes, when the same approved item repeatedly starves the constrained step and a controlled reorder or kitting test increases released output after storage, shelf life, lead time, and cash are considered.
Should forecast demand justify a permanent purchase?
Treat forecast demand separately from committed and repeated demand. Test a reversible option first and examine the downside if expected orders arrive late or at a different product mix.
Practical takeaway
For one product family, log eight representative shifts and run three small challenges: equipment, inventory, and labor. Ask each option to remove the same measured wait and increase acceptable output. The first investment should be the one that survives the test—not the one that sounds most like growth.
