The calculator's new chart led us to extend the value model
The value model has always guarded against counting the same dollar twice: capping how far a group of improvements can move any one measure, dampening what can pile up on any one finance line, and de-conflicting improvement programs that share the same work. This week we extended it to avoid more forms of double counting, and to size each claim more closely to what it can reach.
Much of the credit goes to the chart we added to the calculator two weeks ago. Once you can follow the value from a use case through its measures to the P&L lines it lands on, things a single total never showed are plain to see. In its first days it pointed us to a food plant credited with warranty repairs it does not have, and to the stockout and supplier corrective action refinements we told you about last week. It also showed several related measures meeting on the same line. That was already on our list, but seeing it in the chart is what made us commit to the work, and it is the heart of this week's release.
Related measures pay once. One machine failure shows up as a failure event, as minutes of downtime and as a repair bill: three views of one improvement. The model now maps how 194 measures connect, across 1,417 relationships: which one drives a result, which one is the result, and which is the activity behind it. When a use case claims several of them on the same finance line, that line pays once, through the measure that drives it. A use case keeps the value of an activity when it claims no outcome for it.
Each claim is sized by what it can reach. A measure rarely touches a whole finance line, so each link now records the share it reaches. The model also now separates time freed from money saved. Freed labor time counts at 60%, because a saved hour reaches the P&L only when overtime falls, a leaver is not replaced or hours are cut. Recovered capacity counts at 50%, because freed time becomes product only at the constraint. Those two rates are our judgment rather than measurements, and you can set your own on a baseline.
Improvements on one loss level off. Two improvement programs aimed at the same loss cannot each remove all of it. The model's limits now sit on the loss itself, across 36 kinds of loss, so improvements aimed at the same one curve toward a limit, and revenue growth curves toward the cap in your scenario. Keep stacking work on one problem and the value levels off, the way it does on a real plant floor. Every use case in the catalog at once, at a sold-out $250 million automotive plant, now comes to about $36 million a year.
Costs follow the industry more closely. Not every cost line exists everywhere. Food and beverage plants no longer carry warranty repairs, warranty parts or field service labor, and the other costs in those accounts rise to fill the gap.
What this means for your numbers
Most estimates are lower than they were last week. Added up one use case at a time at a $250 million automotive plant, values are about 27% lower when the plant cannot sell more and about 31% lower when it is sold out.
The change is not even, and use cases that claimed one improvement through several measures move the most. Predictive and condition-based maintenance, an example we have used here, goes from about $1.5 million to about $560,000 a year at that plant, and from about $3.5 million to about $970,000 sold out. Some rise: Inventory Accuracy & Visibility at a $250 million food and beverage plant goes from about $672,000 to about $716,000.
Figures in earlier posts came from the earlier model and will not match what you see now. Saved improvement programs have been recalculated, so what you see in the platform is current.
A value estimate gets used in a room with your finance team. These extensions are there so it holds up when someone pushes on it.
What is this use case worth? It now says so on the page
Every use case on mfgusecases.com now shows a typical value right under its description, as a range: "Typically $500K to $1M a year at a $250M plant." It is a directional estimate of annual EBITDA at typical maturity, with no credit for extra sales, and it names the value model behind it.
See the details opens that use case in the calculator. A free account gives you the exact figure, your own industry and revenue, and where the money lands.
The calculator, closer to your plant
Your own numbers. If your company has a saved baseline, the calculator now opens on it and says so, and one click takes you back to industry-typical. If it does not, the calculator starts on your company's industry.
A point and a percent, spelled out. Every figure now says which move it is and where it lands: "1 point: 92% to 93%. 1% relative: 92% to 92.92%." On a KPI that is already high, those are noticeably different moves, and you should not have to guess which one you are looking at.
A link for every KPI and use case. Each now has its own address, such as /calculator/kpi/first-pass-yield, and the link carries your industry, revenue and capacity state. Send a colleague exactly what you are looking at. Older links still work.
Much faster. The page used to take about eight seconds to open. It now opens in under one.
Corrections
One point of a KPI is now priced at your industry's own level. First Pass Yield, OEE, Forecast Accuracy and Rolled Throughput Yield are affected. One point of yield is worth more at 96% than at 72%, because it removes a larger share of what is left. At a sold-out $250 million plant, a point of First Pass Yield goes from about $713,000 to about $1.3 million in packaging, and from about $944,000 to about $183,000 in biologics. Other KPIs do not change with this.
A point has to be possible. Where one point would run past zero, such as premium freight already under 1% of revenue, the calculator now shows 1% relative instead.
Where to start
Open the calculator and pick a use case you have been weighing for next year. The chart under the answer now shows the value after all of the above, so it is a good moment to look again at which lines your case really depends on. If you have a saved baseline, check your own numbers first.
Calculator figures are for an industry-typical plant of the chosen size and industry, at typical maturity. Your own baseline will give different numbers.