Polaroid Case Study

BACKDROP Polaroid is manufacturer of photographic equipment, accessories and associate items utilise in instant photography. The organization was divided into two master(prenominal) divisions The Consumer Photography Division and the Technical and Industrial Division with each of these divisions change around 40% of Polaroids r even outues of $ 1. 3 billion in 1984. The fraternity produced two main types of films 1. The peel a variance film which required the drug user to physically pull the film pop of the camera and, 2. The integral film, which came out of the camera automatically.The integral films were construct in the R2 building at the Waltham mummy site. The operations at R2 included production of winding-sheet metal rises, pods, flexible cartridges and plastic end caps and then assembled these into film cartridges. R2 ran three shifts, five age a week, employing approximately 900 workers out of which 700 were part time. QUALITY AND PROCESS falsify PROCEDURES A T R2 All films were vetted by the gauge Control division out front being released into the market. The QC procedure included sampling of 15 finished cartridges (each containing 10 frames) out of e truly atomic reactor of 5000 cartridges.If the sampled cartridges contained flees in excess of allowable limits, the lot was held and further scrutinying was done. Additional testing usually led to reworking, or rejection of a portion or all of the lot. Subsequent lots were the subjected to even more rigorous testing by change magnitude the sample size tested. Quality checks were not the sole responsibility of the QC incision. The operators usually sampled around 32 samples out of every lot. If the measurements went against the knowledge of the operator, the sample was rejected.After regale curb was initiated in R2 in the late 1970s, swear out engineering technicians were made obligated for gathering data and qualification rough analyses. PROBLEMS WITH EXISTING QUALITY instr uction Since the testing of cartridges was destructive, it resulted in sampled scrap. This, along with the product that failed acceptance sampling resulted in $3. 28 million in 1984. Another issue was that sampling did cipher to improve quality, it only improved the AOQ. In fact, due to the large production and low defect pass judgment, if the production and quality visualise sampling were halved, the surpass defectives would be 0. 3% of production. On the other hand, increasing the AOQ further would film to prohibitively high costs due to enlarged sampling. The sampling adjoin employed was overly inaccurate. Time was spent on trying to slim beta or consumer risk. Cartridges which were inspected and passed were sent back to production to be repackaged. solely the handling of these cartridges itself increased the chances of their developing defects which resulted in a vicious beat of tests and retests and did not contribute to improving quality significantly while increasi ng costs considerably.To avoid losing production, operators often salted boxes. Operators did not record oft collected data and if they were in doubt, they would pass the grammatical constituent on to the QC Department believing that they would be able to detect the defect and reject the component if the defect was serious enough. Tweaking machines was an accepted practice in the plant. The intention of the exploit was to enable machines running and different speeds and chromosomal mutations to produce at their maximum capacity. The QC department did not focus on defects that were normally notice by consumers.For example, the come nearly strict auditors tested for excess reagent by flipping the film over proper(a) after exposing it, a defect that would not be noticed by a consumer. These stringent auditors averaged about 10% defectives. The conditions on a lower floor which the tests were simulated were also out of sync with current market realities. External customers of ten used cameras which did not function precisely to spec, whereas the QC Department used perfect cameras to test the film. This precluded the possibility of finding defects which would occur with imprecisely functioning cameras. GREENLIGHTThe project objective was quality monitoring costs reduction while at the similar time improving the quality of the product. The improvements in quality ascendancy processes were centre along with reducing the number of samples. The plan consisted of three distinct elements 1. statistical process adjudge would be adapted as processes in control and capable of producing within specifications would produce more consistent quality. 2. Production operators would be given the process control tools that the process engineering technicians had been using and in conjunction with sampling would be expected to subscribe disposition decisions themselves. 3.Quality control auditors would concentrate on training operators and operationalizing specificati ons on their new products. The statistical process control system involving both acceptance sampling and automatise process control was to be implemented. SPC involved testing for productions within a pre-specified kitchen range. If the production went beyond the range, the production process had to be shut cut out tending was to be called to perform guardianship and recalibration. As a part of the process, the operators were to take six ergodic measurements of a process characteristic during the quarrel of their shift and then plot the remember measured value.This led to a drastic reduction in the number of samples tested and consequently the scrapping costs. The primordial problem in this project was the estimation of the central level and the control limits. Initially, the Quality Control auditors helped the operators in plotting the ranges and the operators protocol was to immediately shut mess the machines and call for help whenever, the characteristic crossed the spe cified range. Moreover, eight straightforward mean values lying into the upper or lower zones near the control limits, or consistently upward trends were to be investigated by care as well.The idea behind the project was to cut down the defect and testing losses. However, the idea backfired when the average defects detection by auditors shot up to 10% from 1% while at the operator level, it halved to about 0. 5%. Another problem was the lack of trust betwixt the auditors and the operators. Standardized maintenance procedures also met with a lot of resistance as they were seen as making the whole maintenance process impersonalized and bureaucratic. The operators believed that they could obtain better results by tweaking the machines.At the same time, operators refused to come out of the maximize output mindset and unploughed adjusting the machines for increased output. Also, the operators were sampling and testing more units than they were recording and adjusted the machines on t he basis of the unrecorded defects. The nature of defects also changed. The variability in the kinds of defects detected increased, as the defects recorded by the auditors were markedly different from the defects recorded by the operators. ANALYSIS The purpose of inspection is to determine the level to which the product manufactured conforms to the specifications.Control charts and run tests are used for process control with the objective being to identify the causes of assignable variation, and to leave the system alone if the variation is random and the process is under control. The data given in render 5 was used to calculate the means and ranges of the variables (pod weight and finger height) and the control limits for them were calculated. These let been plotted on control charts. Pod Weight both the X bar and the R chart show that the process is in control, and that the process is capable. The variation present is random variation. Although the X bar chart shows that the process is in control, the last four readings may indicate a trend if further values move towards the lower control limit. Also, between the 16th and 28th readings, there are making of trends. The R chart shows that though the values of R lie within the control limits the range variation is high. Also, the behaviour of the readings is erratic which is a reason for investigation. riffle aloofness The X bar graph shows that the process is out of control very often, signifying that an assignable cause of variation may be present. The values in the R chart are within the control limits. Thus, although the process mean is out of control, the process variability is in control. Other Analysis The random sample of defects from Exhibit 4 is tabulated below. Operator Defects Auditor Defects Excess Reagent 4 11 Excess Flash on box 2 2 Negative sheet defect 3 2 Positive sheet defect 3 3 Double feed 3 3 Frame feed failure 2 9 Damaged spring 3 3 Malformed box 1 3 lean reagent 1 4 Misalignme nt 1 3 Marginal lamination 1 2 Dirt from assembly 0 5 After Greenlight was initiated, the number of defects report by operators has halved from 1% to 0. % while those reported by auditors has increased from 1% to 10%. This may be due to the fact that the operators are not recording all the defective samples which they are using to adjust their machines. Also, since the auditors sprightliness that asking the operators to be incharge of the quality is like handing over the henhouse to the foxes, near of them may have shifted to stringent checking of the cartridges which would explain the jump from 1% rejects to 10% rejects, which was the level of rejects which only the stringent auditors had earlier. There is some evidence for both the preceding(prenominal) points.The tweaking of the machines by the operators may explain why so umpteen readings are out of the control limits, though the machine should have undergone maintenance and calibration as soon as the first reading was outd oor(a) the control limits, which explains why the auditors are finding many more rejects due to the feed than the operators. Also, the auditors are finding more rejects due to the reagent, although the process is under control. This may be due to stringent checking. Another indication of stringent checking is that cartridges are being rejected due to their having dirt which has been attributed to assembly. RECOMMENDATIONS Control measures postulate to be incorporated at the injection molding machines in hostelry to minimize defect rates, and defects need to be prioritized, to help in context of use control limits and the ratings on the quality of products. The operators need to realize that the process downriver is the customer, and they need to shutdown the machine for maintenance as soon as the process goes out of control rather than waiting for the machine to jump-start producing defective pieces. Polaroid can carry out a market seek exercise on consumers, to determine wh ich attributes need compliance from the customers point of view.It will also need to establish the technical specification limits for various components. These will need to build into a 6-sigma process to increase quality by improving the processes and reduce variation in outputs. The people, specially the top management, need to be convinced about the effectiveness of process control, which doesnt have any problem with the quality apart from above observations. Proper documentation of all the procedures and processes should be assured, in order to nutrition people focused on quality once defect rates drop significantly below 1%.This documentation should be reachable to all concerned people and they should be instructed unambiguously to adhere to the norms. automated methods for data collection need to be adopted, like the ones mentioned in the case, since the operators have proved to be unreliable. The investment is not large enough to make a serious dent in the companys bu tt end line, and should be considered. A better and more comprehensive training fabric needs to be introduced to train the workers and supervisors in basic statistics and the application to process control The high-volume driven mindset of the people needs to be changed, and an melodic phrase needs to be built which engenders mutual trust between operators and auditors. extension stress Statistical Process Control Measurements Pod Weight (grams) Sample Number Day Shift 1 2 3 4 5 6 Mean drop 3-Aug A 2. 800 2. 799 2. 760 2. 802 2. 805 2. 803 2. 795 0. 045 B 2. 750 2. 820 2. 850 2. 740 2. 850 2. 790 2. 800 0. one hundred ten C 2. 768 2. 807 2. 807 2. 804 2. 804 2. 803 2. 799 0. 039 4-Aug A 2. 841 2. 802 2. 802 2. 806 2. 807 2. 807 2. 811 0. 039 B 2. 801 2. 770 2. 833 2. 770 2. 840 2. 741 2. 93 0. 099 C 2. 778 2. 807 2. 804 2. 804 2. 803 2. 804 2. 800 0. 029 5-Aug A 2. 760 2. 804 2. 804 2. 806 2. 805 2. 806 2. 798 0. 046 B 2. 829 2. 804 2. 805 2. 806 2. 807 2. 807 2. 810 0. 025 C 2. 741 2. 850 2. 744 2. 766 2. 767 2. 808 2. 779 0. 109 6-Aug A 2. 814 2. 804 2. 803 2. 805 2. 807 2. 804 2. 806 0. 011 B 2. 787 2. 802 2. 805 2. 804 2. 805 2. 804 2. 801 0. 018 C 2. 766 2. 805 2. 804 2. 802 2. 804 2. 806 2. 798 0. 040 7-Aug A 2. 774 2. 801 2. 805 2. 805 2. 805 2. 804 2. 799 0. 031 B 2. 770 2. 801 2. 833 2. 770 2. 840 2. 741 2. 793 0. 099 C 2. 832 2. 836 2. 794 2. 843 2. 13 2. 743 2. 810 0. 100 10-Aug A 2. 829 2. 846 2. 760 2. 854 2. 817 2. 805 2. 819 0. 094 B 2. 850 2. 804 2. 805 2. 806 2. 807 2. 807 2. 813 0. 046 C 2. 803 2. 803 2. 773 2. 837 2. 808 2. 808 2. 805 0. 064 11-Aug A 2. 815 2. 804 2. 803 2. 804 2. 803 2. 802 2. 805 0. 013 B 2. 782 2. 806 2. 806 2. 804 2. 803 2. 802 2. 801 0. 024 C 2. 779 2. 807 2. 808 2. 803 2. 803 2. 803 2. 801 0. 029 12-Aug A 2. 815 2. 815 2. 803 2. 864 2. 834 2. 803 2. 822 0. 061 B 2. 846 2. 854 2. 760 2. 829 2. 817 2. 805 2. 819 0. 094 C 2. 767 2. 804 2. 834 2. 803 2. 803 2. 803 2. 802 0. 067 13-Aug A 2. 850 2. 04 2. 804 2. 804 2. 804 2. 804 2. 812 0. 046 B 2. 810 2. 820 2. 814 2. 794 2. 798 2. 787 2. 804 0. 033 C 2. 850 2. 820 2. 750 2. 740 2. 850 2. 790 2. 800 0. one hundred ten 14-Aug A 2. 750 2. 765 2. 850 2. 760 2. 790 2. 840 2. 793 0. 100 B 2. 830 2. 770 2. 848 2. 760 2. 750 2. 830 2. 798 0. 098 C 2. 740 2. 770 2. 833 2. 770 2. 840 2. 800 2. 792 0. 100 17-Aug A 2. 753 2. 807 2. 805 2. 804 2. 802 2. 804 2. 796 0. 054 B 2. 851 2. 751 2. 752 2. 773 2. 849 2. 806 2. 797 0. 100 C 2. 845 2. 804 2. 803 2. 806 2. 805 2. 806 2. 812 0. 042 18-Aug A 2. 844 2. 777 2. 754 2. 791 2. 833 2. 811 2. 802 0. 90 B 2. 806 2. 839 2. 805 2. 804 2. 850 2. 740 2. 807 0. 110 C 2. 849 2. 801 2. 804 2. 762 2. 814 2. 791 2. 804 0. 087 19-Aug A 2. 820 2. 793 2. 812 2. 833 2. 853 2. 812 2. 821 0. 060 B 2. 790 2. 780 2. 764 2. 843 2. 843 2. 818 2. 806 0. 079 C 2. 850 2. 806 2. 805 2. 814 2. 807 2. 807 2. 815 0. 045 20-Aug A 2. 767 2. 831 2. 808 2. 793 2. 836 2. 811 2. 808 0. 069 B 2. 833 2. 825 2. 793 2. 813 2. 823 2. 766 2. 809 0. 06 7 C 2. 824 2. 799 2. 790 2. 764 2. 817 2. 805 2. 800 0. 060 21-Aug A 2. 778 2. 775 2. 799 2. 805 2. 833 2. 772 2. 794 0. 061 B 2. 801 2. 832 2. 758 2. 759 2. 773 2. 14 2. 790 0. 074 C 2. 770 2. 787 2. 744 2. 766 2. 807 2. 803 2. 780 0. 063 Average 2. 8025 0. 0640 UCL for mean = 2. 8332 UCL for Range = 0. 1280 LCL for mean = 2. 7718 LCL for Range = 0. 0000 Sample Statistical Process Control Measurements Finger Height (mm) Sample Number Day Shift 1 2 3 4 5 6 Mean Range 3-Aug A 1. 90 1. 95 1. 94 2. 00 2. 05 2. 16 2. 00 0. 26 B 2. 15 2. 17 2. 11 2. 13 2. 02 2. 03 2. 10 0. 15 C 1. 73 1. 90 2. 07 1. 89 1. 76 1. 88 1. 87 0. 34 4-Aug A 2. 30 2. 41 2. 54 2. 37 2. 32 2. 16 2. 35 0. 38 B 2. 28 2. 16 2. 19 2. 08 2. 25 2. 24 2. 20 0. 20 C 1. 92 2. 24 2. 1 1. 89 1. 88 2. 17 2. 04 0. 36 5-Aug A 2. 39 2. 28 2. 10 2. 36 2. 54 2. 25 2. 32 0. 44 B 2. 11 2. 21 2. 24 2. 21 2. 17 2. 24 2. 20 0. 13 C 1. 89 1. 90 1. 73 2. 07 1. 89 1. 76 1. 87 0. 34 6-Aug A 2. 51 2. 25 2. 08 2. 35 2. 29 2. 32 2. 30 0. 43 B 2. 22 2. 19 2. 22 2. 24 2. 01 2. 23 2. 19 0. 23 C 1. 89 1. 90 1. 78 2. 07 1. 89 1. 76 1. 88 0. 31 7-Aug A 1. 95 2. 07 2. 25 1. 95 2. 11 2. 16 2. 08 0. 30 B 2. 08 2. 03 2. 27 2. 23 2. 24 2. 13 2. 16 0. 24 C 2. 31 1. 90 1. 86 1. 91 1. 89 1. 87 1. 96 0. 45 10-Aug A 2. 23 2. 25 2. 21 1. 89 2. 15 2. 11 2. 14 0. 36 B 2. 23 2. 21 2. 05 2. 19 2. 7 2. 16 2. 15 0. 18 C 1. 73 2. 00 1. 79 1. 75 1. 84 1. 74 1. 81 0. 27 11-Aug A 2. 21 2. 11 2. 21 2. 44 2. 17 2. 30 2. 24 0. 33 B 2. 17 2. 19 2. 15 2. 04 2. 07 2. 22 2. 14 0. 18 C 2. 01 1. 90 1. 90 1. 81 2. 06 1. 89 1. 93 0. 25 12-Aug A 2. 08 2. 19 2. 28 2. 29 2. 21 2. 45 2. 25 0. 37 B 1. 93 2. 09 1. 90 1. 95 2. 04 2. 09 2. 00 0. 19 C 1. 84 2. 12 1. 90 1. 89 2. 01 1. 75 1. 92 0. 37 13-Aug A 2. 23 2. 01 2. 25 2. 11 2. 39 2. 15 2. 19 0. 38 B 2. 19 2. 22 2. 18 2. 15 2. 23 2. 04 2. 17 0. 19 C 1. 96 2. 05 2. 16 1. 87 2. 13 1. 90 2. 01 0. 29 14-Aug A 2. 27 2. 00 2. 06 1. 97 2. 13 2. 05