Lead-time reduction: shortening the clock the customer actually experiences

The office segments are longer than anyone believes

Take a sample of completed orders and reconstruct each timeline from the customer's clock: order received, order entered, technical query raised and answered, material ordered, material received, work released, work started, work finished, inspected, packed, despatched. The gaps that surprise people are almost always administrative, such as an order waiting for a credit check, a drawing waiting for approval, a requisition waiting for a signature, or a completed job waiting for a collection day. None appear in any production report, and all of them are usually easier to remove than shop-floor time.

A quoted lead time defends itself

When delivery performance comes under pressure, the reliable fix is to quote longer. It works, the performance figure recovers, and the underlying problem becomes invisible while the business slowly loses the orders that needed a shorter answer. The padding then becomes the standard, planners schedule to it, and work sits waiting because there is time in hand. Reducing the quoted figure therefore has to be done deliberately and slightly ahead of the improvement, with somebody watching delivery closely, or the padding will absorb every gain the plant makes. A safer sequence is to shorten the promise for one defined group of products first and watch what happens to the rest.

Batch size is the lever nobody wants to pull

Large batches lengthen lead time twice over, because the batch itself takes longer to finish and it occupies the machine while other work queues behind it. Cutting a batch in half shortens both effects, which is why batch size usually dominates any single operational improvement in its effect on elapsed time. It is resisted because it means more changeovers, worse-looking unit costs and more handling, and because the people who feel those costs are not the people who feel the lead time. Making that trade explicit, and improving changeover so it costs less, is the whole argument.

Engineering and procurement sit inside the clock

For made-to-order and engineered products, the largest single block is often between order receipt and the point where the shop has drawings and materials. Standard components, pre-approved material specifications, an agreed drawing turnaround, framework agreements for long-lead items and a rule about when engineering may still change a design all compress that block. None of it is production work, all of it is manufacturing lead time, and it usually needs a director-level owner because it crosses departments that do not report to operations. Where the same technical query recurs across many orders, the delay is a specification problem being managed as a scheduling one.

Measure the spread, not the average

An average conceals what customers experience, which is the tail. Two plants with identical averages behave very differently if one delivers within a narrow band and the other is occasionally very late. Record elapsed time for every order, look at the spread, and investigate the slowest cases individually rather than dismissing them as outliers, because that is where the mechanism becomes visible. Measure from the customer's order date rather than from works order release, since the release point is under the plant's control and moving it flatters the figure without helping anybody.

Frequently asked questions

Is a shorter lead time always worth pursuing?

Not universally. Where customers plan well ahead and value price over responsiveness, a long stable lead time can be perfectly satisfactory, and shortening it may cost more in changeovers and capacity than it earns. The case is strongest where orders are won or lost on delivery, where forecasts are unreliable so a shorter clock reduces the need to guess, or where the current lead time forces the business to hold stock it repeatedly writes off.

How do we shorten lead time without holding more stock?

Attack the waiting rather than covering it. Smaller batches, work released only when the shop can start it, fewer queue points, faster administrative steps and shorter changeovers all reduce elapsed time without inventory. Stock is the alternative answer and it is legitimate where the customer's expectation is shorter than the shortest achievable process time, but reaching for it first removes the very pressure that would have exposed the delays.

Why did our lead time get worse after we won more work?

Because queueing grows sharply as a resource approaches full loading, and the effect is not proportional to the extra volume. A comfortably loaded plant absorbs an extra order with little delay; the same plant near its limit absorbs it into a queue that lengthens everything else. This is why lead time deteriorates suddenly rather than gradually, and why the answer is usually to control the release of work and protect the constraint rather than instructing everybody to work faster.