Configuration for Cowards

Basics of configuration December 5, 2001

Since 1975 Feldman Engineering Corp. has provided technical support in factory layout, manufacturing methods and computer systems. Visit their website or contact Howard Feldman at

“I’d love to buy it – But can I have it with Shaker style doors, cherry veneer, a darker finish and an inch shorter?” If you can say; “Yes” you either are a one man shop, use a configurator, or a candidate for a nervous breakdown.

You may have a configurator and not know it – Many cut-list and cabinet design programs are based on simple configuration logic. This article discusses the basics of configuration and how advanced applications make it a lot more than a shop floor tool. It is the cornerstone of mixed mode manufacturing, - the equipment and systems that facilitate the efficient manufacture of custom(ized) products in a mass production environment. However, configuration’s greatest value to most companies is as a marketing concept that can be a key profit builder when it controls order entry.

Ancient history: Henry Ford thought he could live without a configurator. He offered cars in “any color you want as long as it is black”. Chevrolet became a major player by offering customers a choice of colors. Lazy but smart engineers decided that it didn’t make much sense to create a different bill of material for each color and simply appended a color suffix code to the car’s model # (A2S[BLU] or A2S[GRN]). The practical limit to colors in the factory were the number of guns and tanks the sprayer could handle but the ultimate limitation was the ability to control finished goods inventory.

It really became interesting when someone asked “Why can’t I have a different color upholstery than gray?” Thus began the first level of configuration: options and features. The car’s identification documents (travelers or route sheets) became more than just information to the workers with a spray gun. The A2S[BLU][RED] now required inventories of subassemblies and purchased parts of both gray and red fabric and trim. An “option or unit value” was added to every inventory item to designate the color sensitivity of every component of the car 0 = no color, 1= body color, 2= trim color. Brake pads and similar parts had no color value but seats, door panels, etc. had to be ordered and ready at the line in each color.

It wasn’t too hard for the assembly-line worker, he had bins of various colored parts and he read the identity tag on the car and matched the color information up with the suffix of the part he was putting on the line. Inventory control at the line was simple – bin min/max (a predecessor of the popular Japanese name inventory reserve schemes) but the real sweat in inventory control was upstream where the rule was never, ever, put anything on the assembly line that you couldn’t complete. Cabinets without doors or chairs without cushions may not seem as big a deal (as a yard full of partially built cars) because they can be stacked awaiting components. But think, your labor supply isn’t completely elastic (nor is your cash flow). Wouldn’t it have been better to build, ship and invoice complete products than build more work in process inventory?

Yes, color options (suffixes) made life easy for engineering as only one bill of material was necessary for a car model. The combination of ten body colors and five interior colors would have required fifty bills (10 x 5) for every model car and five sets of bills for every interior component. Think of the potential for error in doing the same task fifty or more times: think of the needless work to make even a simple update. Eliminating the proliferation of bills of materials avoids chaos in engineering but of greater value is that management can make product rules automatically enforced at order entry to maximize revenue (upcharges for better fabrics and veneers), prevent component inventory proliferation by focusing customer choices ( stock laminate colors vs. long-lead time specials) and control finished goods stock keeping units by restricting choices of inventoried products. The real genius was not the one who decided that offering color choices was a way to maximize sales but the one who realized that you could charge more for special color and trim combinations. At the car factory green bodies would only be offered with beige or gray interiors, red cars were available only in the 2 door model and white cars were $50 more. This is the second level of configuration: rules based option selection.

Setting rules in advance rapidly resolves oddball requests at order entry thus avoiding delays and the need to meet with engineering and inventory control for every non-standard order. The interaction amongst the color options and between the options and the product number are the essence of rule based bills of materials. Of course, you could do this all manually calculating cost one bill of material at a time (different color paints, fabrics and veneers can drastically change costs) and then individually price each product variation. Because you have better things to do on weekends, employ computer system rule based architecture to allow you to make price changes based upon cost, inventory and marketing considerations. “All white cars with blue interiors $200 off this week” – Ford has done this.

The third level of configuration is the most important one for wood products manufacturing: size control. Product and component rules based upon size are no-brainers: fabric width and panel sizes limit construction and/or help the system automatically select alternatives. (Rules for tabletop cores could include limits such as these: maximum width: fir ply 48”; MDF 60”; gum ply 72”.) Depending on the system’s capabilities it would either automatically substitute materials or reject the order if it was out of bounds. Component size control is extremely important for architectural applications and “systems” furniture. The simplest example is a hutch. If your standard construction uses 1/32” laminate faces and the customer selects a special laminate which is available only 1/16” thick, applying this laminate to the two uprights increases the overall width 1/8”. To make it fit the base unit, the shelves and back must be cut 1/8” shorter. The same situation often exists with various thickness and profiles of edge banding. This is not a problem if you work in a 2x4 environment where the customer expects up to a ½” less and sloppy size control. But when you promise an installer 6’ table tops it better be 72” – not 72-1/4 or 71-1/2”.

Level four of configuration is: labor utilization. It is essential for all manufacturing systems to not only measure labor time but to track setup time and waste. This is the basic information for machine loading (accurate delivery projections), labor utilization and realistic costing. The configurator adjusts your process requirements depending on the options selected and the upstream operations and materials. In the tabletop example the 72” wide tops can only be laminated manually and require machining in your oversized bed router. Setup and process time also drive the configurator –and can select your double end tenoner for long runs of tabletops and the router for short runs.

Projecting yield and waste is an important feature of any system used in cabinet, furniture or millwork manufacturing. These are not material losses - other than handling damage there are no yield losses in materials. They can only be incurred in a labor process but different materials will have different factors and require different processes. (Setups can also have a predicted waste factor.) The system must accurately estimate projected losses and then working back from the desired quantity of finished products calculate realistic requirements for materials and process time.

Product size customization is the fifth level of configuration. A leap of faith is to have only one product for each design. This is fairly easy to do with chairs, (which typically are not modified to suit the “Goldilocks” syndrome of too little, too big, too tall) but how about tables? Your catalog could offer: “Style X” table 1-1/8” laminate top, particle board core, banded edges and tubular legs. "Have it your way”! Rules would probably include:

Top SurfaceMaterialAny laminate (hpl)
Top Surface$$7.50 upcharge for non-stocked hpl
Top Surface$15% upcharge for solid color hplHigher scrap rate
Due DateDeliveryAdd 5 days for non-stocked hpl
LengthMaximum120"Cold press/router size
Length$>82"=$9.20 upcharge no-sag brace channel required
HeightLimits24" 26" 30" 36"Legs are purchased parts
Height$36"=$10.50 upcharge
Leg ColorChooseGrey, Black or ChromeLegs are purchased parts
Leg Color$Chrome=$15.00 upcharge
Bulk PackYes/Nono=$3.50 upchargeeffects labor oprations & material

The ultimate leap is to have only one product: table and let the customer configure style, materials, size, shape and type of base. It can and has been done!

How does this all play out in the factory where your nightmare image is a Charlie Chaplin movie of lock-step manufacturing? Will this information overload cause employees to go postal? No! Remember my example of the car factory, the employee just focuses on his single task: he reads the product identification tag and makes one decision at a time. In most cases he will have a work-station schedule (or product traveler) that identifies the process and the options. More complex information can be interpreted directly by a bar code reader on the machine. (Typically labels are created and applied at the panel saw – identifying the job # and part #. The machining center is instructed on the basic machining and the options selected for this part; hole drilling, corner rounding, etc.) The machine doesn’t care or lose time if this part is different than the previous one – neither should your operator! (The only problem is tracking these parts as they move through the factory – but that’s another story.)

Configuration ideally is part of the order entry process (providing instant feedback) but in many large scale business systems it is operated from a stand-alone package. It can either use existing bills as templates (static) or create bills of material directly from rule based logic (dynamic bills). In either case, the bill of material exists only for the one order. Static bills are easier to create and faster to process. Dynamic bills can be used for similar products in varying sizes or to alter the size of an existing product. It is easier to get started with static bills, your staff is probably familiar with their appearance and it is probably easier to modify your system to accept them. They may be all you need, particularly if you aren’t going to change materials or processes based on product size.

In many ways the automobile is a much simpler product than furniture or cabinets – its size can’t be changed by the customer When a customer says the car won’t fit in his garage they steer him to a another product – rather than offer to make it 2” inches shorter. A logical candidate for a static system (without size control)? Yeah, until they started to build trucks of varying lengths and with frame specifications based on payload. The point is first do what ever you need to do to survive today – but don’t block your path to the future!

Let’s get real. I am a strong believer in configuration software for cabinet and furniture companies but I am very concerned with the swing of the industrial pendulum. Our industry was late to embrace computer technology and today many companies who claim to have installed manufacturing business software still have little more than accounting and order entry software – of a quality that is barely adequate for a convenience food store. On the other hand there are companies who have spent more in the last five years on computerization than they have on manufacturing equipment. (Unfortunately in both situations, most companies have focused on “when to manufacture” neglecting “what and how to manufacture”.)

Neither extreme is healthy. Companies with "legacy" systems are living on borrowed time. Customers want it their way and now! The world of “ sixteen week cutting cycles” is long gone. However, never forget that it is the product that you sell, not your ability to manipulate data. Unless you are a “dot com” company, survival and long-term profitability demand that you must first have a salable product, the means to efficiently produce it, and a motivated team before starting down the (often, very long) path towards computer integrated manufacturing.

© 2000 by Feldman Engineering Corporation. All rights reserved. Updated May 07, 2001

Since 1975 Feldman Engineering Corp. has provided technical support in factory layout, manufacturing methods and computer systems. Visit their website or contact Howard Feldman at