The Pumpkin Pie Line: quality control (QC), explained with a Thanksgiving bakery

How the game works

You run QC at a bakery in Thanksgiving week. The line bakes 20 batches of 60 pumpkin pies over five days, and your lab has 200 minutes a day. For each batch you choose what to test and on how many pies, then release it, hold it to test more, or scrap it. Each pie can be checked for fill weight, filling solids, center temperature at the pull, firmness, cracking, crust color and weeping.

‍

Four hidden problems

Every week brings four problems you cannot see at first: a puree lot that is short on solids while its certificate of analysis (CoA) says it is fine, an oven that runs a little hotter every batch, a new egg supplier whose eggs set softer, and a line scale that reads heavy. Overbaked custard cracks and weeps. Wet filling weeps and sets soft. A scale that reads heavy makes real pies light while every weight on it looks right.

Your score is the cost of poor quality: testing, scrapped pies, bad pies that reach customers, the complaint and any fixes. Lower is better.

Around day 4 a customer complaint comes in. You trace it to its cause, and a wrong trace costs you twice.

‍

Spot check or a QC plan

In spot-check mode you look at one pie and weigh it, and look at four more if it seems wrong. In QC-plan mode you set a sampling plan, read control charts with run rules, check the scale against a reference weight each morning and test each puree lot when it arrives. The line runs on its own and stops when something needs you.

Control charts flag a trend while every pie is still in spec. Spec limits only speak up once pies are already bad.

Each game uses one of 2,000 simulated weeks, so the oven, the scale, the eggs and the short lot change every time. You can play the same week in the other mode and compare.

‍

Why inspecting harder is not enough

Testing more pies finds bad batches after they are baked. It does not stop the next one. Across 2,000 simulated weeks, testing 3 pies a batch and acting only on out-of-spec results cost $4,970 on average. A QC plan that spent about the same on testing cost $2,812.

Spot checks cost $8,753 on average, 61% of the week’s $14,400 in pie sales. The QC plan cost $2,812, or 20%. The gap is widest in unlucky weeks: the costliest 10% of weeks averaged $12,783 with spot checks and $3,935 with a QC plan.

Control charts caught the drifting oven before the center temperature left spec in 98% of QC-plan weeks. Checking each puree lot in-house stopped the short lot before it was used in every one.

‍

What the simulation does not do

This is a simplified teaching model. The direction of each effect comes from the research below, but no number was fitted to real bakes. The line pulls pies on time, not on temperature. Cooling, packing and transport are fixed. There is one oven, one scale and one filling line, and the oven only drifts hotter. Food safety is out of scope, and net-content rules are simplified to a single underweight limit.

‍

Worked example: week 16

In week 16 the oven starts drifting at batch 5, the scale reads 11.2 g heavy from batch 5, puree lot PU-1414 has 8.35% solids against a CoA of 9.99%, and the new eggs arrive at batch 11.

• Spot check: every pie looked fine on days 1 and 2. Batch 11, made with the short lot and the new eggs, shipped 59 bad pies. The complaint said “weepy and soft in the middle”, the CoA looked fine, so the oven took the blame. Cost of poor quality: $9,196.
• QC plan: the morning scale check caught the scale before batch 5. A run rule flagged the oven at batch 6, and it was fixed before any pie went out of spec. The lot check read 8.24%, so the lot was quarantined before batch 11. An egg check led to a recipe change. Cost of poor quality: $2,437.

Same week, same pies, same luck. The QC plan found each problem at its source.

From the game to Uncountable

• Setting specs: acceptance limits on every test, with pass or fail on every result.
• Watching the trend: chart results over time with SQC run rules and Cpk, and spot a drift against spec limits.
• Tracing a lot: see which batches used a supplier lot, record the supplier’s CoA values next to your own lot tests, and plot product results by lot.
• Acting on it: point AI Alerts at your data, keep equipment calibration records, and run CAPA and deviation workflows set up for your process.

See how it works on the quality control page.

References

‍

• McGee, H. (2004). On Food and Cooking: The Science and Lore of the Kitchen, revised edition. Scribner.
• Figoni, P. (2010). How Baking Works, 3rd edition. Wiley.
• Gossett, P.W., Rizvi, S.S.H. and Baker, R.C. (1984). Quantitative analysis of gelation in egg protein systems. Food Technology 38(5), 67-96.
• Tong, P. et al. (2020). Influence of heat treatment and egg matrix on the physicochemical and allergenic properties of egg custard. Journal of Food Science 85(3). doi.org
• Nishida, E.-B. and Watanabe, Y. (2024). Effect of components and heating on rheological properties of custard. International Dairy Journal 157, 106024. doi.org
• ThermoWorks. Temperature tips to picture perfect pies. blog.thermoworks.com
• America’s Test Kitchen. Pumpkin pie recipe. americastestkitchen.com
• USDA FoodData Central. Pumpkin, canned, without salt, FDC ID 168450.
• USDA Agricultural Marketing Service (2017). United States Standards for Grades of Canned Pumpkin and Canned Squash.
• Purlis, E. (2010). Browning development in bakery products: a review. Journal of Food Engineering 99(3), 239-249.
• Shewhart, W.A. (1931). Economic Control of Quality of Manufactured Product. Van Nostrand.
• Western Electric Company (1956). Statistical Quality Control Handbook.
• Nelson, L.S. (1984). The Shewhart control chart: tests for special causes. Journal of Quality Technology 16(4), 237-239. doi.org
• Montgomery, D.C. (2019). Introduction to Statistical Quality Control, 8th edition. Wiley.
• ISO 7870-2:2023. Control charts, Part 2: Shewhart control charts.
• ANSI/ASQ Z1.4-2003 (R2018). Sampling procedures and tables for inspection by attributes.
• ISO 2859-1:1999. Sampling procedures for inspection by attributes, Part 1.
• Dodge, H.F. and Romig, H.G. (1959). Sampling Inspection Tables, 2nd edition. Wiley.
• AIAG (2010). Measurement Systems Analysis Reference Manual, 4th edition.
• NIST Handbook 133-2026. Checking the Net Contents of Packaged Goods.
• Deming, W.E. (1986). Out of the Crisis. MIT Center for Advanced Engineering Study.
• Feigenbaum, A.V. (1956). Total quality control. Harvard Business Review 34(6), 93-101.
• Juran, J.M. and De Feo, J.A. (2010). Juran’s Quality Handbook, 6th edition. McGraw-Hill.

‍

Disclaimer

The Pumpkin Pie Line is a free teaching tool built on a simplified model of a bakery line. Results are illustrations, not measurements or food safety, quality or professional advice, so do not use the simulated plans to run a real line.

We have taken care over the model, but the tool is provided as is. To the extent the law allows, Uncountable gives no warranty that the results are accurate, complete or suitable for any purpose, and accepts no liability for any loss that results from using them.