Lesson 12: Capacity for the peaks, and keeping it going

Capacity to the budget

A security-products business had built its edge on reliability: delivery within the lead time customers expect. Its customers, banks among them, needed new branches opened in a month rather than the two or three they used to allow, and a late branch costs them revenue, rent without income, first-mover advantage and sometimes a licence. Demand was skewed too: about 70% of it came in the second half of the financial year. The quarterly split used for planning was 22%, 24%, 26% and 28%.

Yet capacity was planned the usual way: in line with the annual budget’s sales. That rests on assumptions:

actual demand equals budgeted demand, not just over the year but month to month;
spikes can be throttled;
temporary peaks can be moved to cover the troughs.

The reality: rate-contract orders must be served within agreed lead times; some items must be held in stock to meet customers’ tolerance; and demand fluctuates around the budget line every month. When capacity equals the plan, rate-contract orders are served late, stocked items go out of stock and are made to order, orders are rescheduled, and material is diverted between branches. The actions taken make the chaos worse.

The direction of the solution:

Plan about 30% protective capacity beyond the budget.
Have each quarter’s capacity in place in the preceding quarter, so the opportunity is caught from its first day.
Put a dedicated team on capacity addition, working on nothing else.

For each product line, the plan named the output per day needed by a given quarter and the actions to get it: a second or third shift on assembly and final fitting (often the cheapest capacity), an additional welding station, automation and fixtures, an exploitation workshop at a vendor’s plant, moving a product to another plant to free a line.

Pause and think: why put next quarter’s capacity in place a quarter early, rather than when the demand arrives?

Capacity takes time to add, and a promise broken early in a peak loses its customers
Because budgets are approved only at the start of each quarter
Because new capacity runs slowly at first, so starting early keeps utilisation figures high

Shifts need people trained; stations need building. If the load signal arrives first and the capacity months later, the edge is gone in between.

Selling into the troughs

The same business used the plant’s new capability to shape demand as well:

A “happy hours” offer in the lean season, to pull demand ahead into the first half.
Decisions on discounts guided by throughput per unit of constraint time and by the front-of-load: no discount when the constraint is full.
A throughput-based incentive for sales instead of one based on the top line.

Keeping the solution alive

Many implementations decay after the consultants leave. The ones that last share a few rules:

Phase it. Phase 1 is usually manual: choke the release first in production; pilot replenishment on a small number of products. The application manager is involved daily, and the phase ends with a formal learning session. Phase 2 brings the full solution and its IT support.
Watch the reds. When red orders exceed about 10% of open orders, start an immediate analysis.
Meet daily, learn monthly. Daily buffer meetings deal with expediting; monthly learning sessions use the buffer record; every three months, a one-day learning session with the key people.
Make it independent of people. Document the processes in detail, train trainers for new workers, and make the training compulsory for new staff.
Keep improving. Tabulate every reason an order turned red, Pareto it monthly, and attack the top one. As the reasons shrink, cut the buffer again.

Final exercise: which fix?

Match each symptom from this course to the extension that removes it.

On-time delivery swings whenever the share of a slow product rises
One order needs ten days of the constraint, and the buffer is twelve
After halving the lead time, a heat-treatment plant is later than before
Two black orders at the dye house; clubbing opens both and finishes neither
Feeder shops drown in set-ups when the release is order by order
Customer orders lose their common parts to stock whenever stock demand is high
Trucks are made on time but not collected, and sales keeps reshuffling
An expensive constraint’s spindles run idle, but filling them breaks the colours

Check your understanding

Answer all five. Get four right to complete the course.

1. Planning capacity exactly to the budget assumes that…

The bottleneck is in the right place for next year’s product mix
Customers will pay for protective capacity
Demand equals the budget month to month, and peaks can be moved
Vendors deliver on time

In reality demand swings around the budget line every month.

2. Why doesn’t protective capacity raise operating expense in proportion?

Because it is paid for by the customers who order more in the peak
Capacity is set by the bottleneck resources; only they need the addition
Because overtime is free
Because it is temporary

Adding to a few resources protects the whole plant’s output.

3. When should a quarter’s capacity be in place?

When the load signal appears
At the start of that quarter
At the start of the year, all at once
In the quarter before

So the opportunity is caught from its first day.

4. Red orders exceed 10% of open orders. The rule is to…

Start an immediate analysis of where and why they are red
Lengthen every buffer by a third straight away, for all products
Wait for the monthly review
Expedite all of them

Too many reds threaten the protection itself.

5. The common pattern behind every extension in this course is…

A new software system for planning and scheduling
More buffer everywhere
An undesirable effect, a conflict, a wrong assumption, and an injection
Going back to local efficiencies wherever the standard solution fails

The standard solution’s assumptions fail somewhere; the Thinking Processes find where, and how to extend it.