Step 2: Retrieve the manufacturing history of the product being quoted
February 15, 2013
Columnist Gerald Davis continues a series of columns that address one of the most important activities related to shop profitability: job estimating. In this second installment, he describes how an estimator can use the shop’s manufacturing history to refine the predictions of future expense.
This edition of Precision Matters continues our detailed examination of estimating as a business process. The January column covered the need to identify the manufacturing processes needed. If the product has been manufactured in your shop before, then the act of retrieving the manufacturing history is a very efficient method of process identification. Even if the exact product hasn’t been made before, something like it has. So it helps if the retrieval process is versatile. We are going to describe a variety of ways that estimators can benefit from retrieving and analyzing data.
For the sake of discussion, we’ve dissected the estimator’s job into a dozen tasks. In reality, these tasks may overlap or occur in different sequences. For example, which do you do first: retrieve the history or evaluate the design to identify needed fabricating processes? In this specific example, it probably doesn’t matter which step is performed first. However, other steps clearly need to happen in the right sequence; for example, identify the availability of resources before committing to a ship date.
As a review, here is a brief outline of how we’ve dissected an estimator’s job:
We would be delighted to hear from you to help us improve this discussion of estimating, as well as to correct errors and omissions.
Computers and computer software are not endorsed here. Our goal is to look at the job functions and responsibilities of a successful estimator. With a clear strategic vision and armed with a good tactical plan, we should be able to maintain a suite of resources to keep our estimating process on target.
One of those resources might be software to automate the estimating process. We admire efficiency. Many excellent job shop estimating computer products are available. If you use some brand of estimating software to boost your efficiency, you’re an example of good behavior. However, the software is viewed in this article as just a tool to improve the productivity of the estimator. To put it another way, the estimator’s job is not to run software; the estimator’s job is to predict the future.
As an established job shop, you have customers with experienced buyers. Those buyers would probably laugh at the notion of fearing an estimator. However, there are inventors and innovators amongst us who have never had their product manufactured before. Not only are estimators perceived as being judgmental, they also are demanding. How does one approach an estimator to achieve the best possible treatment? The short answer is “with clarity,” but let’s take a longer look at the exchange of forms and events between a product developer and a fabricator.
In the context of this discussion, the process begins with initial contact between the inventor and the job shop. Frequently these two entities are businesses rather than just individuals. The initial contact is a good time to commit to a mutual nondisclosure agreement (NDA). The purpose of this NDA is to protect intellectual property for both parties—identifying what is proprietary and why—and it is best settled sooner rather than later.
To obtain a price quote, the inventor presents to the job shop a request for quote (RFQ), which is internally routed to the estimator. The RFQ is more of an “event” than a “form.” The RFQ might come via e-mail, fax, or other media. An example RFQ is shown in Figure 1.
The RFQ provides the estimator with a desired production quantity and schedule as well as details about what is to be fabricated. It usually includes details about desired shipping destinations and payment terms. When it comes to product specification, it is my habit to include a drawing in the form of a PDF with the RFQ as an indelible quality control document; it is the final arbiter in quality disputes. Figure 2 shows an example of such a PDF.
I also provide the fabricator with 3-D CAD data if they want it. The file format varies depending on their preference, sometimes STEP, IGES, STL, or a native CAD format like SLDPRT.
The job shop’s estimator responds to the inventor’s RFQ with a price quote (see Figure 3). Also known variously as a bid or estimate, the price quote commits the job shop to a price, terms, and a schedule.
The inventor might respond to the price quote with a purchase order. This commits the inventor to payment for the completed work. The job shop confirms the receipt of the purchase order and launches whatever internal steps are needed to fabricate the product for on-time delivery.
As the terms of the purchase order are completed, the job shop presents demands for payment in the form of an invoice. The inventor sends payment in response to the job shop’s invoice.
We are not going to try to cover the full scope of business communication. Keep in mind that the nature of the communication depends on the situation. For example, if a problem exists with the fabricated goods received, the inventor might notify the job shop with a defective-material report. The job shop could remedy the situation with a credit memo, basically the opposite of an invoice.
Who needs estimating? Why not just keep track of the time and material expended and bill the customer for the expense? In some manufacturing circumstances, time and material billing is indeed a good solution, particularly for low-volume or one-off work.
Consider a production situation where batches of the same product are needed every other month. Presumably, the customer will be billed the same amount for each batch. This time and material approach might not give the expected result. For example, one batch might be produced by using on-hand remnants, while the next is made with material that had to be bought twice because of scrap. The erratic time and material pricing is likely to drive good customers away, even though the customer is presumably paying the lowest possible price. Thus, we claim that our estimating process is a customer convenience. They like the care and detail put into the process and expect it with every estimate.
Our estimating process is also a great way to monitor our shop’s performance. The idea is to apply basic quality control procedures that are already in place on the shop floor. In a manufacturing cell, we know what the ideal tolerance performance could be—that is, the prediction of capability. As the cell produces work, we measure the parts and record the deviation from ideal. We evaluate the trends and change the process to maintain production within an acceptable tolerance range.
The estimating process predicts what the expenses for a fabrication cycle should be. A strategic goal is to use the production experience as feedback to improve future estimates for producing the same product. A tactical goal is to minimize the labor required to retrieve and compare predictions to actual results.
The product design shown in Figure 4a features two bends, part marking, and a pattern of holes. Let’s suppose that this product is new to our shop. Figure 4b shows a different part design, one the shop knows because it has made it before. Figure 4b also features two bends, part marking, and uses the same material and plating as the part in Figure 4a.
We could use our experience with the part shown in Figure 4b to create a price quote for the part shown in Figure 4a. All we’d have to change is the flat blank size and the tooling used to create the hole pattern. We can use the work order and expense reporting from the previously made part as a starting point and just modify it to produce the new price quote for the new part.
We have established that data retrieval is useful for estimators. Specifically, retrieval helps to identify needed processes. Also, it establishes a method of getting feedback about the quality of each individual price quote. But those two retrieval and analysis functions are distinct from each other. The search for the history of similar products compares the fabricating sequence from batch to batch. The estimator’s performance report compares the work order expense history to the cost estimate. (Our objective in describing a few of the retrieval and analysis activities that are typical during the estimating process is to provide a frame of reference for both establishing and improving our estimating procedures. Your estimating system probably includes reporting and analysis tools that are more detailed and specific than this generalized discussion will cover.)
Of course, not all price quotes turn into purchase orders. Other areas can be explored to improve the speed and accuracy of the estimating process.
For instance, we can look at the failed price quotes to understand why. If our machinery is out-of-date or our internal processes are not competitive, those lost bids are a database that details specifically which capabilities we need to either upgrade or auction. So, we need a way to analyze lost bids. Who got the work? Did we lose it because of price, delivery, reputation, or delay in responding to the RFQ?
Bids: Subject to Change
We started this retrieval discussion by considering repeat batch production of the same item using the same machinery and the same crew—essentially the same cost basis except for inflation. We need a way to analyze past performance on exactly the same build.
In some instances, a metal fabricator may be forced to approach production of the part in a new way even if there is no revision from the previous build. Perhaps the part was made previously when the shop had a broken machine or experienced a material shortage, or the facility has been upgraded since the last time the part was made, which means the job shop doesn’t have to subcontract out operations. A job may look similar from one purchase order to the next, but change is a constant on a job shop floor.
Sometimes the product itself has been changed from the previous build. A change in quantity or lead-time might have a short-term or one-time impact on cost. A change in the design or packing and handling is likely to have a longer-term impact on cost. Noting and reporting the impact of change from prior builds can help the buyer understand why the price changed. Sometimes that is enough to keep the work even while increasing the price.
In many ways, first-time builds of a product may be profoundly similar—in terms of the steps the estimator takes—to a revision change, particularly if the “new” project is a minor variation on other projects in the shop’s history. As we saw in Figures 4a and 4b, an existing project provides a template that provides a starting point for estimating a new project. That can really shorten the time and labor required for producing a price quote.
New projects that require changes to the facility, labor pool, and supplier network represent operational changes to the business. Such projects are outside the scope of routine cost estimating in a job shop.
Data retrieval and analysis is part of an estimator’s job. Some estimators can get by with memorizing all of the production history and expenses. As production volume builds and the variety of projects increases, an estimator who can provide thorough written records can create a competitive advantage for the shop. Automating the retrieval and reporting functions with software can improve the speed of estimating.
There is no substitute for human analysis of the data. Just printing a report is not enough. The estimator gains skill and improves more rapidly by studying and understanding the information. As an example, consider a report that indicates the job actually cost 100 percent more than the estimate predicted. That could have happened because the estimator missed the difference between “$10 per each” and “$10 per M”—the latter indicating price per thousand. The solution might be to price all things internally as “each,” even though purchasing may buy them by the thousand. That represents a procedural change in response to a persistent problem. On the other hand, the cost overrun may have been because of a machinery malfunction or operator error. In that situation, the estimator may have little to change in the quote; the production crew may need to use the report and analysis to upgrade the equipment or retrain the crew.
Gerald Davis uses CAD software to design and develop products for his clients at www.glddesigns.com. He is a former owner and operator of a job shop from 1984 to 2004. At its peak in 2000, the shop had more than 60 employees working multiple shifts. The insights gained from that experience are reflected in this series of articles.
Gerald would love to have you send him your comments and questions. You are not alone, and the problems you face often are shared by others. Share the grief, and perhaps we will all share in the joy of finding answers. Please send your questions and comments to firstname.lastname@example.org.