1) If a VSM is drawn up for one part family, and any of the work-centers in that VSM are shared with another part family, then does the VSM truly represent the value stream? The answer is a resounding �No.� This shortfall can sometimes be ignored. For example, if a boring mill is used 50% of the time for family A, and 50% of the time for family B, then a modification can be made to the map and its available capacity can be annotated to 50%. The problem with doing this modification of capacity is that the �50%� figure will never stay constant. So, the map must represent an average rather than a true value. An �average capacity� on the boring mill will mean that the VSM under-represents capacity half of the time, and over-represents capacity half of the time, and is never really correct.
2) A conventional VSM assumes that a shop is in reasonably steady state. It assumes the amount of inventory between work centers is relatively stable and can be documented and used in calculations. It assumes the takt time is steady and can be used in calculations. In a job shop, the inventory between work-centers is never constant. Therefore, the VSM may reflect the shop condition at the time it was drawn, but the conditions will quickly and drastically change. If a shop were to make Lean improvements based on the map today, what value would those improvements be when the map changes?
3) A conventional VSM assumes a reasonable average for the work content of jobs in queue. Sometimes this is possible; some times an average value can be very misleading. In a job shop, even jobs in the same part family can have widely varying work content. Although the goal of part families is to minimize the part-to-part differences, sometimes the differences cannot be avoided. In this case, the map may inaccurately describe the work content of the jobs in WIP.
4) One of the biggest drawbacks to a conventional VSM is that it assumes that every job flows to at least a few different work centers. What if there is only one work center in the routing? For example, take a piece of bar stock, turn it on a lathe, and ship it. What value would a VSM be if it only depicted one work center? Worse yet, what about a �Project Manufacturing� site like building a satellite where many operations are done at one work center? A VSM would be essentially useless.
5) Many job shops have a lot of backtracking on their routings. For example, a job goes from machining to fabrication to welding to machining to fabrication to cleaning to welding. Or many sub-assemblies are built simultaneously with the master job, and the sub-assemblies are all routed to a couple of work centers in random order. As all this backtracking is depicted on a VSM, it quickly gets confusing. When a VSM gets visually confusing, it loses a lot of its value as a reference tool.