Developing Annual Operational Plans & Capacity Expansion

Most garment factories have an annual plan. Few have a useful one. The usual plan is a single output target for the year, divided by twelve, with a budget attached. It fails for a simple reason: garment production is seasonal, style mix changes constantly, and capacity is set by a bottleneck that the plan never names. A flat plan is wrong in almost every month, and a plan that is wrong every month is ignored by everyone, which is worse than having no plan at all.

A useful annual plan is built differently. It starts with capacity in minutes, because minutes are stable while pieces change with every style. It forecasts demand month by month, because peak and lean months need different responses. It plans headcount and skills against the monthly gaps, because machines without trained operators are not capacity. And it treats expansion as the last step, tested against orders, bottleneck, staffing and payback, not the first step taken when a big order arrives.

This lesson works through one illustrative factory from capacity measurement to an expansion decision, with every number shown and recomputed. The figures are illustrative, chosen for teaching clarity. The method is what transfers to any real factory.

Capacity starts with a count, not a guess. Count lines, count operators per line, count minutes per day, and count working days per month. The illustrative factory has 12 lines with 40 operators each, giving 480 operators. Each operator works 480 minutes per day, so daily capacity is 480 multiplied by 480, which is 230,400 minutes. Over 26 working days, monthly capacity is 230,400 multiplied by 26, which is 5,990,400 minutes.

Converting minutes to pieces needs two more inputs: average garment SMV and realistic efficiency. At an average SMV of 15 minutes and 75 percent efficiency, monthly output is 5,990,400 multiplied by 0.75, divided by 15, which is 299,520 pieces. Annual output at the same rate is 299,520 multiplied by 12, which is 3,594,240 pieces.

The choice of efficiency matters enormously and is where plans are most often inflated. A plan built at 85 percent efficiency when the factory actually runs at 75 percent overstates output by more than 13 percent, and every month will show a shortfall that looks like a production failure but is actually a planning failure. Use the efficiency the factory actually achieved over the last six months on comparable styles, not the efficiency management wishes for. A plan built on honest efficiency can be beaten. A plan built on wishful efficiency can only disappoint.

Style mix deserves the same honesty. Average SMV is only meaningful if the mix is stable. A factory switching from 15-minute shirts to 25-minute jackets loses 40 percent of its piece output on the same minutes, and a plan stated in pieces will show a catastrophic shortfall while the factory is actually performing well. This is why the plan itself should be kept in minutes, with piece conversions shown per month against that months actual mix. Minutes are the truth. Pieces are a translation.

Demand forecasting starts with what is certain: confirmed orders with quantities and ship dates. Add buyer forecasts, marked clearly as forecasts rather than orders, and add the seasonal pattern from past years: which months peak, which months trough, and by how much. A practical seasonal pattern for illustration is peak months at 120 percent of average monthly demand and lean months at 70 percent, with the rest near average. The exact pattern varies by buyer and product, and the factorys own history is the best source.

Compare monthly demand in minutes against monthly capacity in minutes, and name the months with surplus and the months with shortfall before the year starts. A shortfall month known in January can be managed: overtime planned, temporary workers recruited, a bottleneck eased, or a shipment negotiated earlier. A shortfall month discovered in the shortfall month can only be panicked about. The annual plan earns its keep in this comparison, because it converts a year of surprises into a year of scheduled responses.

Peak months deserve a specific warning. Permanent capacity sized for peak demand stands idle for the rest of the year, and idle capacity still costs rent, maintenance and loan interest. The disciplined response to peaks is temporary: planned overtime, temporary operators, or subcontracting where buyers allow it. Permanent expansion should be sized for the level of demand that persists most of the year, with peaks handled temporarily. Factories that expand for peaks spend the next nine months paying for capacity they do not use.

Headcount planning follows directly from the monthly capacity gaps. A shortfall month needs either more minutes, through overtime or extra lines, or more efficiency, through training and better balance. Each option has a lead time, and the plan must state it. Recruitment takes weeks, training a new operator to full productivity takes longer, and training a multi-skilled operator or a supervisor takes longer still. A plan that assumes 80 new operators arrive fully trained on the day machines are installed is not a plan. It is a hope.

Skills planning matters as much as headcount. A line short of one skilled buttonholer is short of capacity even with full headcount, because the bottleneck operation sets line output. The annual plan should name critical skills per line, current coverage, and the training schedule to close gaps before the months when those skills are needed. Cross-training deserves priority in the plan, because a multi-skilled operator who can cover two bottleneck operations is worth more than a headcount number suggests.

Absenteeism and turnover belong in the headcount plan too. If average absenteeism is 6 percent and annual turnover is 24 percent, both illustrative figures that a real factory should replace with its own, then planned headcount must include cover for absence and a recruitment pipeline for leavers. A plan that assumes full attendance every day will show a shortfall every month, and managers will stop believing it. Plan for the factory you have, including its absenteeism, and the plan stays credible.

Expansion is where annual plans cost real money, so it deserves four tests, all of which must pass. First, the bottleneck test: is the stage being expanded actually the constraint? A factory that adds sewing lines when cutting is the bottleneck gains cost, not output. Find the bottleneck first, with data, and expand it. Second, the orders test: are there orders, not hopes, to fill the new capacity at a healthy utilisation for most of the year? Third, the staffing test: can the operators, supervisors and quality staff be recruited and trained before the machines arrive? Fourth, the payback test: does the investment return its cost within a period the factory can survive if demand softens?

Apply the tests to the illustrative factory. Adding 2 lines takes operators from 480 to 560 and monthly capacity from 5,990,400 minutes to 6,988,800 minutes. At 75 percent efficiency and 15 SMV, monthly output rises from 299,520 pieces to 349,440 pieces, an increase of 49,920 pieces per month. The expansion needs 80 operators, 2 supervisors and 2 quality controllers, and 80 machines at an illustrative BDT 85,000 each, which is BDT 6,800,000 of capital before installation, training and working capital.

Now the tests bite. Are there orders for an extra 49,920 pieces per month for most of the year, or only in peak season? Can 80 operators be recruited and trained before installation? Does BDT 6,800,000 of capital return fast enough to survive a slow season? If any answer is no, the expansion should wait, shrink, or be replaced by a cheaper option: rebalancing lines, fixing the bottleneck, cutting absenteeism losses, or improving efficiency. A factory in Gazipur learned this the expensive way: it ordered 4 new lines for a large new programme, discovered cutting and skilled operators were the real constraints, and watched two lines stand idle while rent and interest accumulated. It later raised output 18 percent without any new sewing lines, by fixing the bottleneck. The lesson is general: expand the constraint, not the line count.

A written annual plan needs five parts. One, monthly capacity in minutes and planned output in pieces against that months mix. Two, monthly demand from confirmed orders and forecasts, with certainty marked. Three, headcount and skills by month, including recruitment and training schedules. Four, capital spending with the four expansion tests documented for each item. Five, owners for every line and a monthly review meeting with a fixed agenda: capacity, output, orders, headcount, costs, and the top three variances with actions.

Revision triggers should be agreed when the plan is written, not invented during a crisis. Practical triggers include orders falling 10 percent below forecast for two consecutive months, a key buyer programme delayed by a quarter, or efficiency running 5 points below plan for a month. When a trigger fires, the plan is revised openly: capacity, headcount and spending change together, and the revision is communicated to everyone affected. A plan revised openly stays trusted. A plan quietly abandoned is replaced by rumour.

Keep one page of the plan visible: monthly capacity, monthly planned output, headcount and the expansion trigger, on the factory heads wall. A plan nobody can see is a plan nobody follows, and the wall page forces the monthly conversation that keeps the full plan alive. The annual plan is not a document produced in January and filed. It is a monthly discipline, and factories that practise it spend the year ahead of their market instead of chasing it.