Does the new layout help on every shift?
Two layouts by three shifts, three days of each combination, eighteen numbers. Both factors are real and the interaction is not, which means the layout question and the shift question can be settled separately.
Operations Advanced Statistics free
After you install, this is the model to open.
Layout and Shift: Which One Moves Output?
- In your spreadsheet, click the Sortia icon in the strip of icons down the right-hand edge. No strip? Click the arrow at the bottom-right to open it. You can also use Extensions, then Sortia, then Open Sortia.
- Click Start from a template and put that name in the search box.
- Pick the card with that name and click Load this template. It arrives on a new tab with real numbers already in it.
The answer
- Layout effect
- F 67.9 cell beats line by 26.9 units a shift
- Shift effect
- F 71.2 day 405.2 to night 357.5, a 47.7-unit spread
- Interaction
- F 0.04 p 0.96: the layout gain holds on every shift
Eighteen shifts: two floor layouts, three shift patterns, and three days of every combination. Because each combination was measured more than once, this test can separate three different stories instead of one. Click Run. The layout effect comes back F 67.94 with a p-value of 0.0000028, so the cell layout genuinely outpacks the line layout, by 26.9 units a shift on these means.
The shift effect comes back F 71.19 with a p-value of 0.00000022 and is the larger of the two: day beats swing beats night on both layouts, from 405.2 units at the top to 357.5 at the bottom, a 47.7-unit spread that is nearly twice what the layout is worth. And the interaction comes back F 0.043 with a p-value of 0.958, which is the line most people skip and the one that decides what you do next.
No interaction means the layout advantage is the same size on every shift, so the layout question and the shift question can be settled in separate meetings, and a floor change will not be quietly undone by whichever crew is on it. To see what the opposite looks like, add 30 units to each of the line layout's three night figures, which is what a night crew that has adapted to that layout would produce, and run it again.
The interaction comes back F 9.33 with a p-value of 0.0036, and now the headline layout average is a lie: the cell layout is 25.7 units ahead on days, 28.0 ahead on swings and 3.0 behind on nights. When the interaction is significant, the two headline averages mean nothing on their own and you have to read the six cell means one at a time, which is what the small block of cell means beside the grid is there for.
One setting here can be silently wrong and is worth knowing about. Rows per sample says how many measurement rows each condition has, and it is 3 because each layout was measured on three days per shift. Set it to 2 and the arithmetic still runs, because six rows also divide by two, but the grid is now read as three conditions of two days rather than two of three, and the layout F drops from 67.94 to 5.18.
Nothing warns you, because that answer is perfectly valid for a grid that really was laid out that way. Check that number against how your data was actually collected before you read anything else. What this cannot tell you is why night is slower: a shift label carries no information about staffing, lighting or handover. To use your own floor, keep the same shape, one factor in blocks down the side and the other across the header, and set Rows per sample to your own number of repeats.
The model
It arrives on a tab called Template: Layout by Shift, carrying these columns:
- Day (units/shift)
- Swing (units/shift)
- Night (units/shift)
with the model computed beside the data:
| Cell layout | 418 |
| Line layout | 392.3 |
Once it is in your sheet
- The model arrives with real numbers in it and runs as it stands, so you can press the button first and understand it second.
- Change the numbers to yours. The sheet marks which cells are inputs and which hold formulas, and most labels carry a note explaining the row.
- Press the run button at the bottom of the panel. It is labeled for the tool you are in, and the result lands on its own tab, with a written reading of it beside the figures.
Never used Google Sheets? Start here goes the whole way, in seven steps, and assumes nothing.
Next question
- How often does the line miss its monthly number?Will the Plant Make the Volume?
- How much cash does the bigger buy tie up?Which Season Plan Should the Chain Commit To?
- Are bookings growing, or is that the zigzag?Smooth the Booking Curve
- Can you spot the bad batch before inspection does?Predict the Defect Before It Ships
- Does the new capacity pay if growth stops?Should We Add the Second Line?
- How often does this plan beat the deadline?Will We Be Ready for Peak?
Every model like this one, and the method behind them: Statistics in Google Sheets.
The method behind this one, worked end to end: ANOVA in Google Sheets.