Quoting systems: pricing work you have not done from data you already hold

An estimate is an unproven routing with a price attached

Everything downstream depends on that framing. The quote contains an operation sequence, a run rate per operation, a setup allowance, a material requirement with an allowance for scrap, and a set of rates that convert time into money. If those elements are not structured the same way as a real routing, the comparison between quoted and actual can never be made, and the estimator is working from intuition dressed up in a spreadsheet. Building the estimate in the same shape as the production routing is the single design decision that makes everything else in this area possible.

Where estimating data comes from and why it ages

Rates originate somewhere: a time study, a previous run, an engineer's judgement, or a number inherited from a machine that has since been replaced. Each has a different reliability and none of them stay true. Processes improve and the estimate keeps the old figure, so the quote is high and work is lost. A machine ages or a material changes and the estimate keeps the old figure, so the quote is low and the work is won expensively. Recording the source and vintage of each rate is unglamorous and it is what lets somebody decide which figures need rechecking first.

The feedback loop that almost nobody closes

Comparing quoted content against booked content, part by part and customer by customer, is the mechanism that makes the next quote better. It needs job booking that is honest enough to compare against, which is the usual obstacle: where operators are pushed to book all attended time to jobs, actuals are inflated and the comparison is worthless. Start with a small set of representative parts rather than the whole book, look at setup and run separately because they fail differently, and treat a consistent gap on one process as a routing problem rather than an estimating problem.

Quoting work you have never made

New parts have no history, and this is where estimating systems differ most. Useful approaches include finding the nearest previously made part and adjusting, decomposing the geometry into features with known process content, and building the estimate from first principles with each assumption recorded as a separate line. The last of these matters more than it sounds: an assumption register attached to the quote lets you explain a price to a customer, revisit it when the drawing changes, and see afterwards which assumption was the one that broke. Unrecorded assumptions become unexplained losses.

Deciding what not to quote

Estimating capacity is finite and most enquiries do not convert. A system that tracks enquiry source, the process content requested, the price offered and the outcome lets you see which customers and which work you win, and at what margin. That is the basis for declining enquiries rather than quoting everything slowly and badly. The pattern worth looking for is work you consistently win on price and consistently lose money on, which usually indicates a process where your estimate is systematically optimistic rather than a customer who negotiates hard.

Frequently asked questions

Should quoting sit inside the production system or alongside it?

Alongside is common and workable provided two links exist: the estimate can be built from the same item, routing and rate data the plant uses, and a won quote can create the order structure without retyping. A standalone estimating tool with its own private rate table drifts from reality within a year and nobody notices until margins fall. The integration that matters is not the quote document; it is shared rates and a path from estimate to routing.

How do we stop estimates drifting from what the shop actually does?

Give the rates an owner and a review trigger rather than an annual sweep. Trigger a recheck whenever a process changes, a machine is replaced, a material specification moves or the actual-versus-quoted gap on a part exceeds a threshold you have agreed internally. Reviewing everything periodically means reviewing nothing carefully, because the volume defeats whoever is assigned. Concentrate on the parts and processes carrying the most quoted volume.

What should we do about work we win and then lose money on?

Separate the causes before acting. Losing money because the estimate was optimistic is an estimating fix. Losing money because the job ran badly is an operations fix. Losing money because the customer changed the specification without a price adjustment is a commercial fix. The data to tell them apart is quoted content versus actual content by operation, plus a record of scope changes. Repricing a customer without knowing which of the three applies usually loses the account and solves nothing.