Value-added time is the fraction of a part's total elapsed time when it is being physically transformed in a way the customer would knowingly pay for. Non-value-added time is everything else: waiting in a queue, moving between cells, sitting in a tote, getting counted, getting inspected. In most plants, value-added time is a thin sliver of the total, often under 10%, and the rest is where lead time actually hides.
This post is about splitting a part's life into time buckets, not about the broader lean idea of value-added versus non-value-added activities. The activity view sorts what people do. The time view sorts every minute a part exists, value-add, business-necessary, and pure waste, and gives you a shop-floor tally sheet to prove where the hours go. It is the measurement layer under value stream mapping.
What is value-added versus non-value-added time?
Value-added time is any minute the part is being changed toward the form the customer ordered: cutting, welding, molding, filling, curing, assembling. Non-value-added time is every minute the part exists but is not being changed that way. The classic test has three parts, and a step is value-added only if it passes all three.
| Test | Value-added step | Non-value-added step |
|---|---|---|
| Does it physically change the product? | Yes, the part is transformed | No, it only moves, waits, or is checked |
| Is the customer willing to pay for it? | Yes, it is in what they ordered | No, they would rather not fund it |
| Is it done right the first time? | Yes, no rework needed | No, inspection and rework repair a miss |
Under that test, moving a pallet 200 feet is non-value-added. So is the tote of parts waiting three hours for the next machine. So is the inspection station, even though it feels productive. The customer did not order an inspection; they ordered a good part. Inspection exists because you cannot yet trust the process to be right the first time.
Why split time into three buckets instead of two?
Because a pure two-way split gets you into arguments you cannot win. Call every inspection, every material move, and every changeover "waste" and the room revolts, some of it is legally or physically required. So the useful model has three buckets, not two.
- Value-added (VA) time. The part is being transformed the way the customer paid for. Protect this; it is the only bucket the customer funds.
- Business-necessary non-value-added (BNVA) time. The part is not being transformed, but the activity is required by law, safety, or current capability: a mandated hold, a regulated inspection, a cure time you cannot yet shorten. Reduce it deliberately over time, but do not delete it today.
- Pure waste (NVA) time. Waiting because the next machine is busy, moving because the layout is bad, counting because nobody trusts the count. This is the free money. Attack it first.
The three-bucket split turns a moral argument ("is inspection waste?") into an engineering one ("how much BNVA time do we have, and what would it take to convert it to VA or eliminate it?"). That is a question a floor team can actually work. It also protects you from the opposite failure, deleting a required hold or a safety check because a spreadsheet flagged it as non-value-added. BNVA is the honest label for "we cannot remove this yet," and yet is the operative word: today's business-necessary cure time becomes tomorrow's target once the process earns the trust to shorten it.
How do you run a time tally on the floor?
Follow one part, timestamp everything, then sort. The goal is a defensible number, not a stopwatch olympiad. Here is the sequence a two-person team can run in a shift.
- Pick one representative part and one route. Choose a high-runner that travels the whole value stream. Do not average across products yet; you want one honest storyline first.
- Physically follow it, do not reconstruct it. Walk with the part or its tote. Reconstructing from the ERP hides the waiting, because systems log transactions, not the hours between them.
- Timestamp every state change. Note the clock time when the part starts a step, finishes it, enters a queue, gets moved, gets inspected. The gaps between timestamps are your non-value-added time.
- Label each segment VA, BNVA, or NVA. Apply the three-part test. When the team disagrees, default to NVA until someone proves the customer would pay for it.
- Sum each bucket and compute the ratio. Total VA time divided by total lead time is your process cycle efficiency. Write all three sums on the sheet so the waste is visible, not just the score.
- Rank the NVA segments by size. One or two queues usually own most of the non-value-added time. That is your first project, not a plant-wide crusade.
- Re-tally after the fix. The same part, the same route, a month later. The ratio moving is the proof; the story of the fix is the coaching.
Watch the cycle time of each step separately from the wait between steps. A team that only measures machine cycle time can shave seconds off a five-minute operation while a three-hour queue sits untouched next to it. The tally makes that trap impossible to miss, because the wait segments show up on the same sheet as the work segments, in the same units.
What does a filled-in tally sheet look like?
Simple: a row per segment, a start and stop time, a duration, and a bucket letter. The sums at the bottom are the whole point. Below is a worked illustration, invented numbers to show the shape, not data from any plant, for a single molded part moving through four operations.
| Segment | Duration | Bucket |
|---|---|---|
| Wait in staging before molding | 95 min | NVA |
| Mold part | 3 min | VA |
| Move to trim cell | 12 min | NVA |
| Queue at trim cell | 140 min | NVA |
| Trim and deflash | 4 min | VA |
| Mandated cure and hold | 60 min | BNVA |
| Final inspection | 6 min | BNVA |
| Totals | 320 min | VA 7 · BNVA 66 · NVA 247 |
Read the totals, not the average. Value-added work here is 7 minutes out of 320, a process cycle efficiency of about 2.2%. The mandated cure and the inspection are business-necessary, so they wait their turn. The two queues, worth 235 minutes between them, are the first and only project worth starting. Nothing about the molding, trimming, or curing steps needs to change to cut this part's lead time nearly in half.
What is process cycle efficiency, and what is a good number?
Process cycle efficiency (PCE) is value-added time divided by total lead time. It is the single number that comes out of a time tally, and it is usually humbling. Unimproved processes commonly land at 5–10%, meaning 90–95% of a part's life adds no value. Well-run traditional plants rarely clear 20%. Lean-improved flows reach 20–25%, and true continuous-flow lines can pass 50%.
Do not chase a benchmark number. The useful comparison is your line against itself over time. A jump from 6% to 12% halves the lead time for the same work content, which shows up as faster delivery, less cash tied up in work-in-process and more room before you need capital. The ratio is a symptom tracker for flow, not a trophy.
By the numbers
The waste is not a rounding error; it is the majority of the clock. Two data points frame it. First, U.S. manufacturers held an inventories-to-shipments ratio of 1.47 in May 2026 on about $962.0 billion of total inventory, per the U.S. Census Bureau's M3 Manufacturers' Shipments, Inventories, and Orders report roughly a month and a half of shipments sitting as stock rather than being worked. Second, the lean literature's repeated finding, echoed across the U.S. EPA's lean methods guidance is that value-added time is a small single-digit-to-low-double-digit percentage of total lead time in most processes before improvement. Both say the same thing: a part spends most of its life waiting, and waiting is where the improvement is.
How does this connect to value stream mapping and throughput?
The time tally is the timeline that runs along the bottom of a value stream map. On a VSM, each process box carries its cycle time, and the sawtooth line underneath separates value-added time from the wait time between boxes. Add up the two and you get lead time; divide and you get PCE. Doing the tally first makes the map honest instead of decorative.
It also connects straight to throughput. Cutting non-value-added time between operations shortens lead time without touching the value-added work at all, and by Little's Law that shorter flow time either frees WIP or raises the rate of finished parts. The two batch companions to this piece, wait-time reduction and work-in-process reduction are the two levers that move the black band in Figure 1.
Where does capture fit?
A time tally done once with a clipboard is a snapshot; the queues grow back. The reason waiting hides is that most systems record transactions, a part scanned in, a part scanned out, not the hours in between. Plants like CLS replaced paper production logging with real-time capture so the gaps between steps are visible as they happen, not reconstructed at month-end. If you want a fast read on what your current losses are worth in dollars before you tally, run your line through a free OEE calculator first, then go follow the part.