The Tart That Costs Thirty-Four Percent

Standard cost per tart:

• Pâte sucrée 0.320 kg × $4.20 = $1.34
• Frangipane 0.180 kg × $9.60 = $1.73
• Pears 3 × $1.05 = $3.15
• Flaked almonds 0.025 kg × $14.00 = $0.35
• Glaze 0.030 kg × $8.00 = $0.24
• Cream 0.050 kg × $6.50 = $0.33
• Total = $7.14, which is 27.5% of $26.00 - essentially on target.

Actual: $18,720 ÷ $26.00 = 720 tarts sold. $6,365 ÷ 720 = $8.84 per tart, i.e. 34.0%. The gap is $1.70 per tart, about $1,205 a month.

That size of gap is almost never ingredient price - price moves are gradual and visible on invoices. It is a yield and portioning variance. The three candidates, in order of what I would measure first:

1. Pear halves. A fourth half to fill the tart is $1.05, i.e. 62% of the entire gap, and nobody weighs pear halves.
2. Dough trim. 320 g per shell, but if the sheeter trims and re-rolls at 20% loss, that is 400 g issued, +$0.34.
3. Frangipane portion. A scoop that dispenses 210 g instead of 180 g is only 30 g, but at $9.60/kg that is +$0.29.

Together those three explain $1.68 of the $1.70. Before touching the recipe, do a yield test: weigh the dough issued against shells produced for one batch, weigh ten scoops of frangipane, and count pear halves against tarts built. Check the trim bin and the staff-room bin too.

1. Cut frangipane by 10%. Saves $0.173 per tart - 0.7 percentage points, taking you from 34% to 33.3%. It also changes the product: frangipane is the structural, moist layer under the pears, so a thinner layer gives a drier, flatter tart and more collapse at the cut. Cost of the fix: a worse tart and the customer notices. Reject as the primary lever.

2. Pre-ground almonds. Ground almond meal is typically a few percent cheaper than whole almonds plus your grinding labour, so perhaps $0.10–$0.15 per tart - again under 1 point. Against that: ground almonds oxidise and go rancid far faster, you inherit whatever freshness the supplier has, and you lose control of the grind texture that gives frangipane its bite. It is a convenience decision, not a food-cost decision. Neutral at best.

3. Raise the price to $29.00. At the current $8.84 actual cost, food cost falls to 30.5%; at the standard $7.14 it falls to 24.6%. It raises revenue by $2,160 a month at current volume, but you may lose units, and critically it does nothing about the leak - the $1.70 overspend scales with every tart you sell, so growth makes it worse.

Recommendation: fix the portioning first. Closing the $1.70 returns you to 27.5% and recovers roughly $1,205 a month without changing price or recipe. If, after that, the shop still needs margin, a modest price rise is defensible on a properly costed card - but decide what a pear tart is: if four pear halves genuinely look and eat better, re-specify the card and price for it, rather than shipping an undocumented over-portion.

The failure is that a cost card existed and nothing ever compared it to reality, so put the comparison on a routine rather than relying on vigilance.

At the bench (stops the variance):

• Written scaling sheets in grams with batch yields, signed by the person who scales and countersigned by a second.
• Fixed portioning tools: a scoop sized to deliver 180 g of frangipane, a pear count written on the build card, dough balls individually weighed before sheeting.
• A stated trim rule - one re-roll maximum, trim weighed and recorded - so dough loss is visible rather than absorbed.
• A daily waste log with a price column, not just a description.

At the desk (detects it):

• A weekly yield test on the tart line: dough issued vs shells produced, tarts built vs pear halves issued.
• A monthly theoretical-vs-actual reconciliation: (units sold × standard cost) compared against issues and purchases, adjusted for inventory movement. Keep the two within 2 points; investigate any gap over $200.
• Re-cost the card whenever two or more invoice lines move more than 5%.

Verification: run the reconciliation for the next three months. If the variance sits at or under 2 points and the yield tests hold, the system is working. If it drifts, look for the largest single contributor first - in this business it is nearly always pears, cream, or dough trim, not the price of almonds.