Shop Operations
How to find and fix bottlenecks in a countertop shop (lean for stone fabricators)
Find the step that limits your countertop shop's output, keep it busy, and use lean tools like WIP limits, batching, 5S and standard work to shorten lead times.
, CEO, Stonify
A bottleneck is the one step whose capacity limits how many jobs the whole shop can finish. In countertop shops it is rarely the bridge saw. It is usually the CNC or edge finishing in a one-machine shop. In a growing shop it is often template drawing and programming, and at high volume it is often the install crews. Find it by counting where jobs pile up and comparing waiting time with working time. Then keep that step busy, feed it only jobs that are ready, and fix scheduling and process problems before buying equipment. Lean manufacturing adds the daily tools for this: seeing waste, limiting work in progress, batching similar work, 5S and standard work.
What a bottleneck is, and why it controls everything
Think of a shop as a chain of steps: template, draw and program, cut, CNC, edge and polish, inspect, install. Each step has its own capacity. The shop as a whole can only finish jobs as fast as its slowest step. That step is the bottleneck, also called the constraint.
This idea sits at the center of the Theory of Constraints, set out by Eliyahu Goldratt in The Goal (1984). It has two practical consequences that surprise most shop owners:
Improving a step that is not the bottleneck does not raise output. If the CNC can finish 16 kitchens a week and install crews can do 20, a third crew adds nothing but cost.
An hour lost at the bottleneck is an hour lost for the whole shop. An hour lost anywhere else is usually absorbed by slack.
So the first job is to know which step is the bottleneck. The second is to protect its time.
Why lead time is mostly waiting
Most shop owners feel that their lead times are long because the shop is busy. Usually it is because jobs are waiting in piles between steps. A standard kitchen needs about 10 to 15 hours of hands-on work in total, yet the usual template-to-install time is 5 to 15 business days. See the worked example in workflow overview.
Little's Law, a basic result from operations research, links the three numbers:
Average lead time = work in progress ÷ throughput
If a shop has 40 jobs between template and install and finishes 10 a week, the average job spends 4 weeks in that stretch. To shorten lead time without working faster, you have to hold fewer jobs in progress. Letting more jobs into the shop "to keep everyone busy" makes lead times longer, not shorter.
How to find your bottleneck
Count the piles
For one week, write down at the same time each day how many jobs wait in front of each step: templates to draw, programs to write, slabs waiting for the saw, pieces waiting for the CNC, pieces waiting for edge work, finished jobs waiting for an install slot. The bottleneck is almost always just after the biggest, most persistent pile.
Compare waiting with working
Take your last 15 to 20 jobs and record the date each one entered and left each stage. For each stage, compare the hands-on time with the days elapsed. Stages where jobs wait days for an hour of work point to the constraint, or to a handoff nobody owns.
Watch for the signs
One station is always running, often with overtime, while the next one waits.
Operators downstream finish early or clean up while waiting for parts.
People walk to the office to ask what to do next.
A pile of finished pieces grows near the door while the install calendar is full.
Check takt time
Takt time is how often you need to finish a job to keep up with demand:
Takt time = available production time ÷ jobs needed in that time
With 480 minutes of CNC time a day and 8 jobs a day to finish, takt is 60 minutes per job. If the CNC averages 75 minutes per job including setup, it cannot keep up. It is the bottleneck, and jobs will queue in front of it.
Common bottlenecks by shop type
These patterns show up often. Your own counts matter more.
Shop profile | Where the bottleneck often sits | Why |
|---|---|---|
Small shop, one CNC, 5 to 8 people | CNC or edge finishing | One machine does cutouts and profiles for every job |
Mid-size shop, two machines, 10 to 15 people | Template processing and programming | One or two people turn every template into a drawing and program |
Shop growing fast | Information flow | Specs, approvals and slab assignments do not keep up with job count |
High volume, 60+ jobs a month | Install crews and scheduling | Fabrication scales faster than trained crews |
The bridge saw is rarely the constraint in modern shops. A saw can usually cut far more slabs a day than the CNC and edge line behind it can finish.
The bottleneck also moves. Fix the CNC queue and the constraint may shift to programming, then to installs. That is normal. The method is a loop, not a one-time fix.
The office bottleneck
A large share of lead time in many shops is spent before any stone is cut: waiting for a drawing, a customer approval, a sink model or a slab decision. If your counts show jobs waiting days between template and first cut while the machines have slack, the constraint is in the office. More machines will not help.
Fixing the bottleneck: the five focusing steps
The Theory of Constraints gives a simple order of work.
Identify the constraint, using the counts above.
Exploit it. Get more output from the constraint as it stands, with no spending:
- Never let it stand idle at breaks or shift change. Stagger lunches.
- Prepare the night before: programs loaded, slabs staged next to the machine, tooling checked, work tickets printed.
- Feed it only released jobs with confirmed specs, so it never cuts something that comes back.
- Batch similar work to cut setups (see below).
- Inspect slabs before they reach it, so it never works a slab with a hidden crack.
Subordinate everything else to it. Other steps work at the constraint's pace. This is called drum-buffer-rope. The constraint sets the pace (the drum). A small stock of ready work waits in front of it, about half a day to a day, so it never starves (the buffer). And new jobs are released into the shop only as fast as the constraint consumes them (the rope). Cutting ahead of the CNC just builds a pile of cut pieces that can chip, get mixed up or wait for days.
Elevate the constraint only after steps 2 and 3. Cross-train a second person to run it. Add a partial second shift. Outsource overflow cutouts or edge work. Buy equipment last. A second-shift operator costs a year's wages. A new CNC costs six figures before installation, power and floor space. Measure the machine's real cutting time against its setup and idle time before you decide. A machine that spends much of its day waiting does not need a twin.
Repeat. Once the constraint moves, start again at step 1.
Lean in a stone shop
Lean manufacturing comes from the Toyota Production System. Its core idea is to deliver what the customer ordered with as little wasted time, motion and material as possible. It fits a countertop shop well, because every job is custom and most lead time is waiting.
The seven wastes, stone-shop edition
Waste | What it looks like in a countertop shop |
|---|---|
Transportation | Slabs moved three or four times before the saw; pieces carried across the shop and back |
Inventory | Piles of cut pieces waiting for the CNC; slabs bought months before they are needed; unsorted remnants |
Motion | Walking the yard to find a slab; hunting for a profile wheel; walking to the office to ask what's next |
Waiting | Templates waiting to be drawn; the CNC idle while a slab is found; crews waiting for a missing piece |
Overproduction | Cutting jobs ahead of what the CNC can take; fabricating before approvals are in |
Over-processing | Re-polishing an edge that was fine; drawing the same job twice because the template was retyped |
Defects | Remakes from template errors, wrong edge or wrong sink cutout; chips from rough handling |
Some lean practitioners add an eighth waste: unused skills, such as the experienced fabricator who is never asked why jobs come back.
Tools that work in a countertop shop
Value-stream map. Draw every step from signed contract to installed job on one page. Under each step, write the hands-on time and the typical wait before it. In a 3 to 5 person shop this takes a few hours and usually shows two or three obvious fixes. The map almost always shows that waiting dwarfs working.
Pull and work-in-progress limits. Set a maximum number of jobs allowed to wait at each station, for example four slabs waiting at the CNC and three at edge work. When a station is full, the step before it stops starting new work and helps elsewhere. This is the practical form of the rope. It feels wrong at first. It is what shortens lead times.
Batching and setup reduction. Group work by material, thickness and edge profile to cut changeovers. Sources and shop reports put each changeover at about 10 to 30 minutes, so on a busy day batching can return an hour or more of machine time. Then attack the setup itself. Keep profile wheels and tooling for the day at the machine, preset and labeled. Standardize the order of tool changes. Do anything that can be done while the machine is running before it stops.
5S. Five steps to organize a work area: Sort (remove what isn't used; one common rule is anything untouched for 90 days leaves the floor), Set in order (a place for everything, marked, near where it's used), Shine (clean, which in a wet shop also means finding leaks and worn parts), Standardize (write down the setup so every shift keeps it) and Sustain (check it regularly). In stone shops, 5S pays off most at the saw, the CNC tool rack, the remnant rack and the install trucks.
Standard work. Write down the one best current way to do repeat tasks: the polishing grit sequence per material, the pre-cut slab check, the loading and protection of a truck, the pre-install call. Variation between operators shows up as uneven finish time and uneven quality. Standard work makes both predictable, and it is how you train new people.
Visual management and a daily standup. A board or screen that shows each job's stage, plus 5 to 10 minutes each morning to look at red flags. Problems caught a day or two early are cheap. Problems caught at the install are not.
Quality at the source. Rework is the worst kind of waste because it uses capacity twice, often at the bottleneck. Build checks where errors start: verify every template against the sale before programming, confirm specs in writing, inspect each slab before it goes on the saw. See remakes and quality and template errors and verification.
Cross-job batching
Cross-job batching means cutting pieces from more than one job on the same slab. A vanity for one customer goes into the offcut from another customer's kitchen in the same color. It raises slab yield and uses fewer slabs. The trade-offs are real. The two jobs' schedules become linked, a delay on one holds the other, and the tracking has to show which piece belongs to which job. It works best for common stocked materials and small pieces that fit in predictable offcuts. See waste and yield and slab layout and nesting.
Fix things in the right order
Scheduling first. Release rules, a production freeze, loading the constraint to about 80 to 85%, batching. These cost nothing and show results within a few weeks.
Standardization second. Checklists, grit sequences, loading procedures, 5S. Low cost. They take one to two months to become habit.
Software third. Job tracking that records stage dates, nesting and inventory. Moderate cost. Most shops need a few weeks to learn it well.
Equipment last. Only when the data shows the machine itself is the constraint after steps 1 to 3. Procurement to full use typically takes months.
Metrics that show whether it's working
Lead time (template to install, or contract to install), tracked weekly.
Work in progress by stage: how many jobs sit in each column.
Utilization of the constraint: actual cutting time as a share of available time. Well under the planned level while jobs wait means a feeding problem, not a machine problem.
First-pass quality and remake rate: jobs that go backward in the flow. A few percent is commonly reported in residential work. There is no published industry benchmark, so track your own trend.
On-time install rate.
See job costing and KPIs for how to build these into a weekly review.
Common mistakes
Buying a machine before measuring where jobs wait.
Treating lean as a one-time cleanup day ("5S twice a year").
Keeping everyone busy by starting more jobs, which raises work in progress and lead time.
Fixing five things at once, so nobody knows which change helped.
Leaving the crew out. The people at the saw know where the time goes.
Ignoring the office, where drawings, approvals and slab decisions often hold jobs for days.
Where software helps
You cannot fix a bottleneck you cannot see, and most shops have no record of when each job entered and left each stage. In Stonify each job's steps are tasks with an assignee (a person or a machine) and dates. The All Tasks report can export template and install dates to measure template-to-install time, and custom views list open projects with their task dates side by side, so jobs that have waited too long stand out. Calendar tracking metrics total square footage or task hours per day for each person or machine, which shows where load piles up. Availability blocks keep maintenance time on the constraint visible. In the drawing's layout step, Auto Nest places the pieces and estimates the slab count in one click, which cuts the hand layout work behind each quote.
FAQ
Is the saw or the CNC usually the bottleneck? In most shops with modern equipment it is the CNC or edge finishing, not the saw. The saw cuts rectangles quickly. The CNC and edge line do the slow work of cutouts, profiles and polish.
How do I know if I need a second CNC? Measure how much of the day the current CNC actually cuts versus waiting for programs, slabs or setups. If it is waiting a lot, fix feeding and setups first. If it is cutting nearly all available hours and jobs still queue in front of it, a second shift or a second machine is justified.
What is the difference between lean and Six Sigma? Lean targets waste and flow speed. Six Sigma targets variation and defects. Most stone shops get more from lean first, because waiting is their biggest loss. Defect reduction methods help once the flow is stable.
Does lean apply to installation? Yes. Standard install kits per truck, a pre-install checklist and readiness call, and confirming every piece is on the truck before leaving are all lean practices.
Can a three-person shop use these ideas? Yes, and it is quicker. Mapping the workflow takes an afternoon. In a small shop the bottleneck is often the owner, who sells, templates and answers every question. Writing down standard work and handing off one role is often the biggest gain.
Related Articles
How a countertop shop's workflow runs, from lead to install
The stages a countertop job passes through, who owns each handoff, what has to be true before a job moves on, and where jobs get stuck.
Countertop production scheduling: how to plan templating, fabrication and installs
How countertop shops schedule templates, cutting, finishing and installs: capacity in hours and square feet, release rules, batching, rush jobs and install dates.
How to lay out a countertop fabrication shop: building size, flow and utilities
Floor space, zones, one-way slab flow, machine clearances, ceilings, floors, power, water and aisles for a stone countertop shop, with a worked layout for a mid-size shop.