Job Shops
Quote the one-off job, find out what it actually cost, and see what the hot job displaces before you promise it.
Work arrives as an RFQ with a drawing attached. Every job is different, and the quote is the bet.
A job shop lives or dies on the quote. A print arrives, somebody decides how long setup takes, what the cycle time is, and what the material costs, and that number is a bet that settles weeks later. Most shops never settle it. The hours land on a time card, the invoice goes out, and nobody compares the estimate to what the job consumed. WorkCell closes that loop as a by-product of running the work: the estimate becomes the routing, operators clock the operation at a terminal, and job cost posts as hours and material move. When the hot job walks in on Thursday, finite capacity scheduling shows which promised dates move before you agree to anything.
A new part gets half a day of setup, a familiar one gets three hours, and the number comes from whoever is holding the print. That is a reasonable way to estimate. What breaks is the absence of feedback: nobody goes back after the job runs to see whether half a day was right. The shop quotes the next ten jobs off the same instinct, and the instinct drifts with the machine mix, the material and the operator. Three months of recorded setup times would settle it, and almost nobody has three months of recorded setup times.
Quoted hours sit in a spreadsheet or the quoting package. Actual hours sit on paper cards in a rack, in a punch clock export, or in the foreman's head. Reconciling them is a manual exercise that happens at year end if it happens at all. The shop learns its overall margin from the accountant and learns nothing about which jobs, which customers, or which operations are eating it.
Machining time is the part everyone estimates. The rest of it is fixturing, soft jaws, measuring, deburring, cleanup and the recut after the first article comes back out of tolerance, and that is the part that quietly consumes the margin. Because none of it is a routing operation, none of it is captured, so the next quote repeats the omission and the shop concludes it is simply bad at machining.
A customer drops off parts that have to run today. Say no and they walk, say yes and the work already promised slips. Either way the spreadsheet with a tab per machine is now wrong, and the only person who knows the real sequence is whoever maintains it. The jobs that moved get discovered late, usually by the customer whose date was quietly sacrificed. The decision itself was never the problem; the shop just had no way to price it in dates before saying yes.
The same bracket comes back a fourth time and the estimator starts over, because last time's actuals were never attached to the part. A shop that has run a job a hundred times should be quoting from its own average, not from memory. Without a record tied to the part and its revision, every repeat order is priced as if it were new work and the shop's best asset goes unused.
Quoting that becomes the job it priced
Estimates are built from operations with setup and run time, material, and outside processing, then converted into the routing the floor runs. There is no re-entry step where a quote is translated into a work order by hand, so the sequence the estimator priced is the sequence the shop executes and the one the actuals come back against. Quoting the same part at three quantities is a change to the amortization rather than three separate estimates.
Job cost posted as the work happens
Every labor entry, material issue and completion posts against the job the moment it is recorded, stamped with the job as its cost object. Margin on an open job is visible mid-run rather than after the invoice, which makes a job going sideways a decision you can still make instead of a number you discover in a quarterly report.
Estimate against actual at the operation, not the job
Variance is reported per operation, so the shop learns that setup ran two hours over while run time was fine, rather than that the job lost money. That is the resolution an estimator needs to fix a standard. Over enough jobs it turns instinct into a number the shop can defend to a customer who asks why the price went up.
Finite capacity scheduling with proposal runs
The scheduler plans against real machine and labor capacity rather than assuming the next operation is always free. Proposal runs let a planner sequence the hot job, see which promised dates move, and either commit the plan or throw it away. The uncomfortable conversation with a customer happens before the commitment instead of after the miss.
Shop floor terminals that capture quantity, scrap and time
Operators clock on to an operation at a terminal, record good and scrapped quantities, and clock off. The costing record, the schedule status and the traceability all come from that single entry, so the floor is not filling in a time card, a traveler and a production log to say the same thing three times.
Routings and bills of material that carry a revision
Multi-level bills and routings are versioned and effective dated, so the fourth run of a part shows the structure it was built to and the actuals that came with it. Engineering change is versioning and effective dating rather than a change-request workflow, which suits a shop where the customer print, not a committee, is the authority.
Quoting and estimating built from operations, material and outside processing.
ProductionFinite capacity scheduling, proposal runs, and operator terminals on the floor.
AccountingJob costing posted as labor and material move, against the job as a cost object.
AnalyticsEstimate against actual by operation, part and customer across every job you run.
Quoting, engineering, inventory, scheduling, shop floor and job costing on one database.
MESOperator terminals capturing quantity, scrap and time at the operation.
APSFinite capacity scheduling for a floor where the mix changes every week.
CPQThe quoting category, for shops whose estimating problem comes before the job.
Precision CNC shops where setup against cycle time decides the margin.
Fabrication ShopsCut, form, weld and finish work with outside processing in the middle.
Tool and Die ShopsLong, one-off builds where the engineering iteration is the job.
ProductionScheduling, work orders and the terminals the floor actually uses.
How do you quote setup time on a part you have never run?
Most shops estimate setup from the part and their offset library, then amortize it across the order quantity. The part that matters is checking it afterward. WorkCell reports actual setup against estimated setup per operation, so the number on the next quote comes from your own history instead of instinct.
How many shops actually track actual hours against the bid price?
Fewer than claim to, because tracking only survives when capturing the hours costs the operator nothing. WorkCell puts it on a shop floor terminal at the operation rather than in a timesheet filled out on Friday. The variance report is then a by-product of running the job rather than a separate chore.
Can I schedule the floor without a spreadsheet tab per machine?
Yes. Finite capacity scheduling plans against the machines and the labor you actually have, and proposal runs let you test a sequence before committing it. When a hot job arrives you see which promised dates move, which is the one thing a tab-per-machine spreadsheet can never tell you.
Does job costing pick up deburring and inspection time?
Only if they exist as operations on the routing, which is the point. Make the deburr, the inspection and the rework real steps and operators clock them like any other. The hours stop disappearing into the machining number, and the next estimate for that part carries them explicitly.
What happens when the same part comes back a year later?
The part, its revision, its bill of material and its routing are already there, along with the actual hours and material from every prior run. Quoting the repeat becomes pricing from your own averages rather than re-estimating from scratch, and the estimator can see which run was the clean one.
Is WorkCell a fit for a ten-person job shop?
It is built for discrete manufacturers from roughly ten to a thousand employees, and high-mix, low-volume build-to-print work is the case it is designed around. A shop that quotes constantly and needs estimate-against-actual feedback is a closer fit than a long-run repetitive producer.
See where WorkCell fits your shop.
Tell us how you quote and schedule today and we will tell you straight whether WorkCell is a fit.