Takt timeTakt time (German Taktzeit — beat or pace)

Takt time is the available production time in a period divided by the customer demand in the same period. It answers a single question: "How often must a finished unit roll off the line to satisfy demand without falling behind and without building inventory?"

If a line is open for 7.5 productive hours per shift (450 minutes after breaks + meeting time are removed) and the customer wants 900 units that shift, takt time is 450 ÷ 900 = 0.5 minutes per unit. The line must finish a unit every 30 seconds. Whether it CAN finish a unit every 30 seconds (cycle time) and whether it actually IS finishing one every 30 seconds (actual cycle time) are separate questions — important ones, but separate.

Takt time = Available production time ÷ Customer demand. The challenge is computing both inputs correctly.

Available production time

Not gross shift length. Subtract: planned breaks, planned meetings + handovers, planned changeovers (if takt is being computed at the line, not at the cell), planned preventive maintenance windows, planned tea / safety meetings. The result is sometimes called Planned Busy Time (PBT) under ISO 22400-1; sometimes called Available Operating Time. The number that goes into takt is the time the line is actually scheduled to produce.

Customer demand

The forecast or order-board commitment for the period the available time covers. Demand should be levelled (heijunka) — averaged over a planning horizon, not the spike of a single order — so the takt does not whipsaw shift to shift. Mixed-model plants compute a separate takt per product family.

Example calculations

The three are routinely confused. They are different things.

Takt is not a static plant-engineering parameter. A modern MES publishes takt to the line in real time, updates it as schedule changes land, and displays it next to actual cycle time on every workstation. The kiosk shows the operator: target = 30 sec, current = 28 sec (green), or current = 36 sec (red, andon-warning).

1. Schedule publishes per-WO target output + planned operating window.
2. MES computes takt = window ÷ output and pushes to every workstation in the cell.
3. Kiosk displays takt next to actual cycle time as a live ratio ("On pace" / "+12 sec / unit slow" / "-3 sec / unit ahead").
4. Pitch board — the wall display showing planned units vs actual units at hourly increments (sometimes called the hour-by-hour board) is computed from takt × hour.
5. Andon thresholds are set as multiples of takt (e.g. raise an andon if actual cycle time exceeds 1.5 × takt for 5 consecutive cycles).
6. OEE Performance is computed as (ideal cycle × good count) ÷ run time — where ideal cycle is derived from the line design, but the gap to takt drives operational decisions like adding overtime.

Takt is most associated with discrete assembly (cars, devices, consumer electronics) because the unit boundary is unambiguous. It applies to process manufacturing too — but the unit definition takes thought.

For batch process manufacturing, takt time at the unit level is meaningless; instead the equivalent concept is average inter-batch time — the average time between batch starts of the same product family. This drives scheduling, equipment availability, cleaning windows and operator staffing. For continuous manufacturing, takt converts to throughput rate (units of mass / time) and ISO 22400-2 §4.3.2.4 applies.

Takt only works as a discipline if demand is levelled. A line running takt = 30 sec / unit over a shift of three products, with one product piled into the first 2 hours and the other two into the last 6, is not running at takt — it is running in batches with a constant pace number stamped on top.

Heijunka (平準化, "levelling") is the discipline of distributing product mix evenly through the shift so the line genuinely runs to takt for every product family at every hour. Tools include the heijunka box (a physical or digital schedule grid that releases work in small, mixed pitches), every-product-every-interval (EPEI) scheduling, and pull-system sequencing.

Regulated plants — pharma, supplements, devices, food — sometimes treat takt as a Lean concept that does not apply to them. They are mistaken. Takt becomes more important in regulated environments, not less, because in-process holds and deviations introduce variability that the schedule must absorb, and only a takt-aware schedule can absorb it intelligently.

1. Schedule realism — knowing takt makes it impossible to commit to a release date that the line cannot achieve; the conversation with QA / supply / commercial becomes evidence-based.
2. Holds and deviations have a quantified cost — a 2-hour hold on a line running at takt = 30 sec / unit costs 240 units of output; the cost is a deviation-record-attached fact, not a feeling.
3. Validation runs — IQ / OQ / PQ runs need to demonstrate that the line can sustain takt under representative conditions; takt is the validation acceptance criterion that maps directly to commercial viability.
4. OEE / 22400 — the Performance term of OEE is sensitive to takt; without an honest takt number the OEE is decorative.
5. Capacity planning vs the regulatory window — radiopharm, sterile, fresh-meat and blood-products plants have hard time windows (decay, sterility hold, shelf-life) within which production must finish. Takt against those windows is the binding constraint, not against generic demand.

Mistake 1 — using gross shift length

Including breaks, meetings, planned changeovers and planned PM in available time inflates takt (makes it look slower than reality). The line then misses targets and no one understands why. Use Planned Busy Time per ISO 22400-1.

Mistake 2 — single takt for a mixed-model line

A line that builds three products of different cycle complexity cannot be managed with a single takt number. Compute a weighted takt or per-product takt; use heijunka to sequence so the line runs to the right takt for each pitch.

Mistake 3 — confusing takt with cycle time

Telling operators "your takt is 30 seconds" when you mean "your cycle time is 30 seconds" causes them to push past the design cycle to chase the demand pace, which causes quality escapes. Display both numbers separately on the kiosk.

Mistake 4 — ignoring the pitch / hour-by-hour board

Takt as a number on a dashboard nobody looks at is theatre. The pitch board — visible at the line, comparing planned vs actual at hourly increments — is what makes takt operational. Without it, takt is engineering trivia.

Mistake 5 — chasing takt past line design capacity

If demand drives takt below cycle time, the answer is not to push operators harder. The answer is capacity investment, line redesign, second shift, or — sometimes — pushing back on commercial commitments. Pretending takt is achievable when it is not is how plants burn out their best operators.

Mistake 6 — not republishing takt when schedule changes

When a WO is added, removed or resequenced, takt for the rest of the day changes. If the kiosk and pitch board are not updated, operators run to a stale target. The MES has to republish takt on every schedule write.

V5 ships takt as a live, schedule-driven number computed per line × cell × product, published to the kiosk and the pitch board, and wired into andon + OEE.

• Schedule-driven — every WO release writes target output + planned operating window; takt = window ÷ output is recomputed for the affected line + cell on every change.
• Kiosk display — every workstation shows takt + actual cycle time as a live ratio (green / amber / red), with a one-line note on the configured pitch interval.
• Pitch board — /app/pitch shows the hour-by-hour planned vs actual board for every running line, computed from takt × elapsed; planned units, actual units, gap, andon events overlaid on the same row.
• Andon thresholds — the andon system fires automatically when actual cycle time exceeds the configured multiple of takt for the configured run of consecutive cycles, surfacing problems before the supervisor notices.
• OEE / 22400 — the Performance term + the throughput KPIs come from the same takt × output stream — no parallel data capture, no reconciliation cycle.
• Mixed-model — per-family takt is computed from the WO mix; heijunka pitch sequencing publishes the right takt to each pitch, not a single misleading average.
• Continuous + batch — for batch process tenants the equivalent inter-batch time is displayed; for continuous lines the throughput-rate equivalent (units/hour or kg/hour) is computed; the kiosk uses the right language per industry profile.
• Mobile-safe — pitch board works on iPhone (≤390 px CSS width) with no horizontal scroll, so the supervisor walking the floor sees the same numbers as the kiosk on the line.

Is takt time the same as cycle time?

No. Takt is the pace the customer requires; cycle time is the pace the line is designed to run at; actual cycle time is the pace it currently runs at. Lean production aligns the three: actual ≈ cycle ≤ takt. Conflating them is the most common pitfall.

Does takt time apply to process / pharma manufacturing?

Yes — at packaging it applies directly (pack / blister / carton is a unit). At API and bulk production the equivalent is average inter-batch time. At continuous manufacturing it converts to throughput rate (kg/hour). The concept of "customer pace as a binding scheduling constraint" applies in every mode.

How do you set takt for a brand-new product with no real demand?

Use the forecast for the launch period and reset every shift / week as actual demand lands. Be explicit that takt is a forecast quantity for the first weeks — operators understand the difference and prefer it to fake precision.

What's a healthy gap between cycle time and takt?

10–20% slack is the rule of thumb — enough to absorb normal variability (small stops, minor quality events, operator changeover) without being so much that the line is structurally underutilised. Less than 5% is fragile; more than 30% is wasteful capacity that should be redeployed or downsized.

How does takt change for shift overlaps and breaks?

It does not change; takt is computed over Planned Busy Time, which already excludes breaks. What you do is design the line to either pause during breaks (most common) or stagger breaks across operators so the line runs continuously — in which case takt is computed over the unbroken operating window.

Is takt useful for a job shop / make-to-order operation?

Less directly, because demand is by definition irregular. The useful adaptation is to compute average takt over a rolling 4-week window for repeating product families, and use it to set staffing and equipment utilisation expectations. For genuinely one-off jobs takt is replaced by individual lead-time commitments.

How does V5 publish takt to the line?

Schedule writes → MES recomputes per line × cell × product → kiosk + pitch board update within seconds. The number is never re-entered, never on a poster, never out of date. Every change to the schedule is immediately visible at the line.