Kanban Boards and Visual Workflow Control in Stone Shops

Most stone shops run on a whiteboard, a spreadsheet and the memory of one very stressed production manager. Jobs are pushed into the shop as soon as they are sold, work in progress piles up between stations, and the answer to where is job 4417 depends entirely on whether the right person is in the building that day. Kanban offers a different model: work is pulled through the shop only when the next station has capacity, and the current state of every job is visible to anyone standing in front of the board.

The word describes a signalling system rather than a piece of software. In its original manufacturing form a kanban was a physical card that travelled with a batch and authorised the next step. For a fabrication shop the useful translation is a wall-mounted board with columns for each stage of production, cards for each job, and hard limits on how many cards may sit in any column at once. It is deliberately low-technology, and that is a feature rather than a compromise, because a board on the wall is seen by everyone whereas a scheduling screen is seen by whoever opens it.

Push, Pull and the Cost of Work in Progress

A push system releases work based on when it arrived or when it was promised. A pull system releases work based on whether the downstream station is ready to receive it. The difference sounds academic until you count the slabs currently standing on carts between the saw and the CNC, each one representing money spent, floor space consumed and a delivery date that is not moving closer. Every slab in that queue also ages: templates get revised, customers change sinks, and material that has been cut for a job that has stalled cannot easily be repurposed.

Work in progress has a direct relationship with lead time. Where a station processes a steady number of jobs per day, the time a job waits in queue rises in proportion to how many jobs are already waiting. Halving the queue roughly halves the wait. That is why limiting work in progress, rather than working faster, is usually the fastest route to shorter lead times in a shop that is already reasonably productive.

Limiting work in progress also exposes problems instead of hiding them. When queues are deep, a saw that is down for two hours is invisible because everyone downstream has plenty of buffered work. When queues are shallow, the same outage stops the shop and everyone notices. That sounds worse and is in fact better, because a problem that is visible gets fixed while a problem that is buffered persists for years and quietly consumes capacity.

The counter-argument every production manager raises is that stone work is too variable for a pull system. Some jobs are a single vanity, others are a forty-piece commercial package. That variability is real, and it is handled by sizing cards in work content rather than in job count, or by running separate swim lanes for small and large work. It is not a reason to abandon the approach; it is a reason to design the board around the actual work mix.

Designing a Board That Matches Your Shop

Map the real process, not the ideal one

Walk a job through the shop and write down every place it stops. Most shops discover two or three stages they never counted, typically inspection, rework and staging for delivery. Those stages belong on the board because they are where jobs actually wait. A board built from the process as it is described in a quality manual will not match the process as it runs, and the crew will stop trusting it within a fortnight.

Set explicit limits and enforce them

Every column gets a number written at the top. When the column is full, no new card enters until one leaves. The first week of this is uncomfortable because it makes the constraint visible, and the natural reaction is to raise the limit. Resist that for at least a month. The discomfort is the diagnostic information you were previously paying for and not receiving.

Make the card carry the information

A useful card shows the job number, customer, material, piece count, promise date and any special condition such as a mitred edge or a sintered slab that needs different tooling. Colour coding by material family lets a fabricator see at a glance whether the day is mostly granite or mostly engineered stone, which matters for tooling changeover. Keep it to what fits legibly on a card; a board that requires reading a paragraph is a board nobody reads.

Put it where the work is

The board belongs on the shop floor, at eye level, in a spot people naturally pass. In an office it becomes a management artefact. On the floor it becomes a shared reference that fabricators use to plan their own next task without asking anyone.

Running the Board Day to Day

The board needs a rhythm. A ten-minute stand-up at the same time each morning, in front of the board, with the whole production team, is the standard pattern. Move left to right through the columns, discuss anything blocked or ageing, and finish with the day's commitments. Avoid turning it into a status meeting where everyone reports; the board already shows status, so the conversation should be about exceptions only.

Ageing markers make delay visible. A small dot added to a card each day it stays in the same column turns a tidy-looking board into an honest one. Three dots on a card in the polishing column tells a story that no verbal update will. Some shops use a simple date written in the corner; the mechanism matters less than the fact that time in column is visible.

Resist the urge to expedite. Every expedited job displaces others and adds a switching cost that nobody measures. Where genuine emergencies occur, use a single, permanently limited expedite lane, typically one card wide, so the shop can absorb a true emergency without the practice spreading. If the expedite lane is always occupied, the promise dates being given by the sales team are not achievable and that is the real problem to fix.

Connect the board to the digital system rather than replacing it. Job data still lives in the shop's management software, and the board is a physical view of a subset of it. Reconciling once a day, usually by the production coordinator during the morning setup, keeps the two aligned without creating double entry. Where a shop has an existing scheduling screen, a large monitor showing the same columns can work, though physical cards remain easier for a crew to move and mark.

Measure a small number of things. Lead time from template to delivery, throughput per week, and percentage of jobs delivered on the committed date are enough to tell whether the system is helping. Tracking twenty metrics guarantees that none of them get acted upon. Post the three chosen numbers next to the board and update them weekly.

Sustaining the System and Avoiding Common Failures

The most frequent failure is quiet abandonment. The board works for six weeks, a busy month arrives, the limits get suspended so that everything can be pushed through, and it never recovers. The defence is to treat the limits as a policy with a named owner rather than as a suggestion, and to schedule a monthly review where any permanent change to a limit is discussed explicitly rather than drifting.

The second failure is a board that describes the process nobody follows. Shops change: a new CNC arrives, an edge polishing line is outsourced, a second shift starts. The board must change with them. Build a habit of reviewing the column structure quarterly, and be willing to delete a column that no longer represents a real stopping point.

The third failure is using the board as a performance monitoring tool for individuals. The moment cards become evidence in a disciplinary conversation, people start moving cards to look good rather than to reflect reality, and the board's information value collapses. Keep it firmly at the level of the work rather than the worker.

Maintenance planning integrates naturally once the board exists. Because the board shows when a station will be starved, it also shows the best window for a blade change, a spindle service or a tooling changeover. Shops that plan preventive maintenance into visible gaps rather than reacting to breakdowns typically find the change pays for the whole exercise on its own.

Over a longer horizon the board becomes a capacity planning instrument. Consistent bottlenecks at a single column, month after month, are an equipment or staffing case that can be made with data rather than with argument. That is a materially stronger position than asking for a new machine because the shop feels busy, and it tends to survive scrutiny from whoever signs the cheque.

None of this requires software, consultants or a factory-scale operation. A four-foot magnetic whiteboard, a set of coloured cards and the discipline to hold the limits will change how a ten-person shop feels within a month, and the improvement in delivery reliability is usually visible to customers well before it shows up in the accounts.

Multi-site operations add a wrinkle worth planning for. A shop with a separate template crew, an off-site polishing partner or a second location for large-format work has handoffs that a single board cannot represent honestly. The usual answer is a board per location plus a shared handoff column that appears as the last column on one board and the first column on the next. Cards physically travel with the work, or a photograph of the card is sent ahead so the receiving site can plan.

Seasonality is another factor stone shops feel acutely. Kitchen renovation demand swings across the year in most markets, and a board tuned for a busy spring will feel absurdly loose in a quiet January. Rather than rewriting limits every few weeks, set them for the normal season and treat sustained periods above or below as a signal to adjust staffing or take on different work, such as commercial packages that tolerate longer lead times and fill the trough.

Training new staff becomes noticeably easier once a board exists. A new fabricator can see the whole production sequence on one wall, understand where their station sits, and grasp who they receive from and hand to without a formal orientation. Shops that have run boards for a while report that the board does much of the induction work by itself, which matters in a trade where experienced fabricators are hard to recruit and internal promotion is the realistic path.

Customer communication improves as a side effect. When the production coordinator can look at a wall and say a job is in polishing and scheduled for delivery Thursday, the answer is specific and it is true. Vague reassurance is the main source of the follow-up calls that consume front-office time, and those calls fall away quickly once accurate status is available on demand.

Finally, keep an eye on what the board cannot see. Quality problems that are caught after delivery, material that was ordered incorrectly weeks earlier, and estimating errors baked into the original quote all sit outside the production flow. A board improves the part of the business it covers and can create false confidence about the parts it does not, so pair it with a simple defect log and a periodic review of jobs that lost money despite flowing smoothly.

Predictable production depends on predictable tooling supply. Explore diamond blades and consumables that keep stations running between planned changeovers, and browse the full Dynamic Stone Tools range for the equipment behind each column on your board.