Future State Map: The 8 Questions Step by Step, With a Worked Example

Rother and Shook reduce value stream design to eight questions that you answer in a fixed order: first customer and takt, then flow and pull, then scheduling, and last the improvements you need. This lesson walks through each question on the steering bracket example and checks every answer with numbers. The result is a production lead time of 4.0 days instead of 11.3 and roughly two thirds less inventory. The design guidelines themselves are covered in the lesson on future-state design guidelines; this lesson is about the procedure.

Starting point: the current state

Steering bracket, left and right hand, 920 pcs per day, two shifts with 920 minutes of available time in total. Five processes: stamping (1 s, changeover 60 min, batch 1,000, uptime 85%, EPEI 3 days), welding 1 (39 s), welding 2 (46 s, uptime 80%), assembly (62 s, 2% scrap) and shipping. In between: coils for 5 days, stamped parts for 2.5 days, a FIFO lane with 40 assemblies, 1,600 welded parts (1.74 days) and finished goods for 2 days.

The timeline gives a production lead time of 11.3 days against 153 s of processing time. Inventory excluding purchased parts: 4,600 + 2,300 + 40 + 1,600 + 1,840 = 10,380 pcs (coils counted as piece equivalents). The bottleneck is assembly at 63.3 s effective cycle time; welding 2 is just below it at 57.5 s.

Questions 1 and 2: takt time and finished goods

Question 1: What is the takt time? It is the yardstick for every later decision. Set the planned cycle time of a process slightly below it so that downtime and scrap still fit.

Question 2: Will you build to a finished goods supermarket or directly to shipping? The steering bracket is a standard part with daily, fairly even customer releases. So a finished goods supermarket stays, but it is limited: one day of release cycle plus half a day of safety against fluctuations in the releases, so 1.5 days. For customer-specific parts you would build directly to shipping instead.

Questions 3 and 4: Where can it flow, where do you need a supermarket?

Question 3: Where can you use continuous flow? Look for neighboring processes with cycle times close to takt and no significant changeover. Welding 1, welding 2 and assembly meet these conditions: 39 + 46 + 62 = 147 s of work content. They are combined into one cell; the FIFO lane and the 1,600 welded parts disappear.

Stamping does not fit into the cell: 1 s cycle time against 60 s takt, a long changeover, and the press supplies both variants in batches. In the cell it would spend almost all its time waiting.

Question 4: Where do you need supermarket pull systems? Wherever a process cannot be linked by flow. That is why a stamped-parts supermarket sits after stamping, and the cell pulls from it with withdrawal kanban; each withdrawal triggers replenishment through the kanban post. A coil supermarket sits in front of stamping, and the steel mill will refill it daily instead of twice a week.

Questions 5 and 6: pacemaker and product mix

Question 5: At what single point will you schedule production? Only one process receives customer demand: the pacemaker process. Here it is the cell. Upstream of it, supermarkets are replenished by pull; downstream, parts flow to the finished goods supermarket and to shipping. The weekly schedules for stamping, welding 1 and welding 2 go away; production control sends the daily release to the cell only.

Question 6: How will you level the production mix at the pacemaker? Instead of building all left-hand brackets first and then all right-hand ones, the cell switches variants at short intervals. This works because the cell has practically no changeover. The gain shows upstream: stamping sees even consumption of both variants and can run smaller batches, and the stamped-parts supermarket needs only one day.

• Demand per variant: 460 pcs per day left, 460 pcs right.
• Leveled: for example, alternate one container left, one container right.
• Not leveled: only one variant for a whole day, so the supermarket has to buffer the other variant’s demand for days.

Questions 7 and 8: pitch and required improvements

Question 7: What increment of work will you consistently release at the pacemaker? The pitch is the time for one shipping container. With 20 steering brackets per container and a 60 s takt, that is 20 minutes. Production control or the material handler releases one container every 20 minutes, for example through a heijunka box. A deviation from takt then shows up after 20 minutes at the latest, not at the end of the shift.

Question 8: What process improvements are necessary for the design to work? Every assumption in the previous answers becomes an action with an owner and a due date; in the editor it is a kaizen burst placed where it takes effect.

• Stamping changeover from 60 to 10 minutes (SMED). The batch drops from 1,000 to 300 pcs, and changeover per piece stays small at 600 s ÷ 300 = 2 s. Effective cycle time at stamping: 1 ÷ 0.85 + 2 = 3.2 s instead of 4.8 s.
• Uptime in the cell to 95% (TPM). Today welding 2 has only 80%; at 80% the cell would run at 56 ÷ 0.8 ÷ 0.99 = 70.7 s, above takt.
• Reduce assembly scrap from 2% to 1%.
• Set up the cell with three operators and redistribute the work content (balance each operator to 56 s at most).
• Agree on daily coil deliveries instead of twice a week.

Result: current and future state compared

Production lead time drops from 11.3 to 4.0 days: coils 1.5 + stamped parts 1 + finished goods 1.5 days. Inventory excluding purchased parts drops from 10,380 to 3,680 pcs (1,380 + 920 + 1,380), a reduction of 6,700 pcs or roughly two thirds. The processing time of the parts hardly changes; almost all of the gain is waiting time.

Be careful with the value-added ratio. The timeline adds up cycle times along the main path. In the future state the cell counts with its 56 s cycle time, not with its 147 s of work content. Processing time on the timeline therefore falls from 153 s to 62 s (1 + 56 + 5), and the value-added ratio rises only from 0.025% to 0.028%. Compare current and future state by lead time and inventory, not by value-added ratio.

• Production lead time: 11.3 → 4.0 days (−65%).
• Inventory excluding purchased parts: 10,380 → 3,680 pcs (−65%).
• Bottleneck: assembly at 63.3 s, above takt → cell at 59.5 s, below takt.
• Scheduling points: three (kanban, weekly schedule, heijunka) → one (pacemaker cell).

Implementation: plan in loops

A future state is not introduced all at once. Divide the value stream into loops: the pacemaker loop (cell to customer) and the upstream loops (stamping with its supermarket, supplier with the coil supermarket). You usually start with the pacemaker loop because it sets the takt for all the others.

Each loop gets a value stream plan for the year: objective, actions, owners, due dates and a metric to track progress. In the editor, set the map’s state to future, enter the actions as kaizen bursts and quantify the gain with the current/future comparison. Use the simulation to check whether the supermarkets hold up when cycle times vary and breakdowns occur.

Frequently asked questions

Keep learning