Touch time is the hands-on time actually spent working on a unit, cutting, assembling, packing. Lead time is the whole elapsed clock from order or release to delivery. Touch time is a slice of lead time; everything else is the unit waiting. Divide one by the other and you get flow efficiency, usually a small number.
The reason to put these two side by side is that their ratio settles an argument most plants keep having: is our lead time long because the work is slow, or because the work is waiting? Touch time answers the first half, lead time the whole thing, and the gap between them is pure waiting. This post defines both, works the ratio, explains why it is almost always tiny, and separates it cleanly from the more familiar lead-time-versus-cycle-time comparison.
What is touch time, and what is lead time?
Touch time is the sum of the value-adding work performed on a unit, the minutes a person or machine is physically changing the product toward what the customer ordered. If you filmed the unit and cut out every frame where nothing was being done to it, what remains is touch time. It is close to, but a bit stricter than, total processing time, because it excludes setup, inspection, and rework. A useful discipline is to agree, per product, exactly which steps count as touch and which do not, and to write it down. Two engineers who draw the line differently will report different touch times for the same part, and the ratio only earns trust when the definition is fixed and shared.
Lead time is the door-to-door clock the customer experiences: order placed, then everything that happens before, during, and after production, ending when the product is delivered. It wraps order handling, queue time before the first operation, all the in-process waiting, and shipping. Lead time is what you promise; touch time is what you actually do. The distance between them is where delivery dates live or die.
How do touch time and lead time differ?
They differ in what they count and, more importantly, in what they reveal. Touch time is small, stable, and about the work. Lead time is large, variable, and about the wait. One is an engineering number; the other is a customer experience.
| Dimension | Touch time | Lead time |
|---|---|---|
| Counts | Hands-on value-adding work only | All elapsed time, order to delivery |
| Typical size | Minutes to hours | Days to weeks |
| Mostly made of | Actual processing | Waiting in queues |
| Improved by | Faster methods, better tooling | Less waiting, smaller batches, less WIP |
| Who feels it | The process engineer | The customer |
The trap is assuming they move together. Shave 10% off touch time and, if touch time is a tenth of lead time, you have cut lead time by about 1%. The customer will not notice. Attack the waiting instead and lead time can drop by a third with the machines untouched, because the waiting was the lead time all along.
What is the touch-time-to-lead-time ratio?
That ratio is flow efficiency: touch time divided by lead time, expressed as a percent. A unit with 30 minutes of touch time and a 3-day (say 24 working-hour) lead time has a flow efficiency of about 2%. The number sounds alarming the first time a plant measures it, and then it becomes the single most useful diagnostic they own, because it points straight at where the improvement is.
A low flow efficiency is not a sign of lazy workers or slow machines, it is a sign of a process built around waiting. And that is good news, because waiting is cheaper to remove than capacity is to add. A plant obsessed with touch time buys faster equipment and wonders why lead time barely budges. A plant that watches flow efficiency goes after the queues and cuts lead time without spending on steel. For the deeper treatment of the denominator, see flow efficiency; here the point is that the ratio is the gut check.
Why is the ratio usually so small?
Because waiting is manufactured by ordinary, well-intentioned choices, and it stacks up fast. Batches make every unit wait for its neighbors. Shared equipment makes work queue for its turn. Handoffs between departments add a wait at every boundary. Safety stock and released-too-early work pile work-in-process that each unit must sit behind. None of these feels wasteful in the moment; together they turn a few hours of work into a few days of lead time.
In many lean assessments, flow efficiency comes in under 25%, and office or job-shop processes often sit in the single digits, the value-add work is real but rare against the elapsed clock. The exact figure matters less than the direction it points: a small ratio says stop optimizing the work and start deleting the waits. That is why the ratio is a better first question than "how do we speed up the line?"
How is this different from lead time vs cycle time?
Lead time versus cycle time compares two durations at the system level: how long a unit takes end to end versus how often a finished unit comes off the line. Touch time versus lead time compares work against wait to produce a ratio. One tells you the rhythm of output; the other tells you how much of the clock is value-add. They answer different questions and are easy to confuse.
Put simply: lead time vs cycle time is about pace and promises, can I make a unit often enough to meet demand, and what do I quote the customer. Touch time vs lead time is about waste, how much of my lead time is actually doing anything. A plant can have a healthy cycle time and a dismal flow efficiency at the same time: units come off steadily, yet each one spent days waiting to get through. You need both lenses, but only the touch-to-lead ratio tells you the lead time is mostly air.
How do you improve the touch-to-lead ratio?
Raise the ratio by shrinking the denominator, not the numerator. Lead time is mostly waiting, so that is where the leverage is. Work this order:
- Map the waits. Walk one unit's real path and timestamp every queue. You cannot delete waiting you have not made visible; a value-stream map or a simple travel log exposes it fast.
- Cut batch sizes. Smaller transfer batches mean each unit waits for fewer neighbors before moving. This is usually the biggest single lever on flow efficiency.
- Cap work-in-process. Release less work into the process. Less WIP means shorter queues, and by Little's Law shorter lead time at the same output.
- Kill handoffs and approvals. Every boundary between people or departments adds a wait. Combine steps, co-locate, or pre-authorize so units do not sit in an inbox.
- Attack the constraint's reliability. A bottleneck that stops injects machine downtime that ripples into every downstream queue, stretching lead time far beyond the stop itself.
- Re-measure the ratio. Recompute flow efficiency after each change. If lead time fell but touch time held, you removed waste, not work, exactly the goal.
What do the numbers say about the room to improve?
A low touch-to-lead ratio in a plant with idle capacity is almost always a flow problem, not a capacity one, and most plants have that idle capacity. Per the Federal Reserve's G.17 Industrial Production and Capacity Utilization release U.S. manufacturing ran near 75.8% capacity utilization in spring 2026, roughly 2.4 points below its 1972–2025 long-run average. When a quarter of capacity sits unused and lead times are still long, the lead time is waiting, not working, which is precisely what a small flow efficiency tells you, and precisely the waste that costs nothing but discipline to remove. Longer view of the same journey lives in our guide to throughput and the plant KPI scorecard.
Where does tracking fit in?
Flow efficiency needs two honest numbers: real touch time from the work content, and real lead time from timestamps at the start and end of a unit's journey. Both die when they are reconstructed from memory at shift end, because that is exactly when the queue detail gets rounded away. Capturing release and completion times where the work happens, the shift from paper travelers to real-time data that plants like CLS made, is what turns flow efficiency from a one-off kaizen measurement into a number you can watch. See how the pieces connect in Harmony's real-time operational layer.