Design to Cost · Evaluate
Estimate Confidence Grading
Grade every estimate A/B/C: quoted, analogous or guessed — and never mix them in one decision.
- Time30 min
- FormatSolo
- StageEvaluate
Estimate Confidence Grading: what it is and why it works
Estimate Confidence Grading attaches a quality label to every cost number before it enters a decision. Grade A means the figure comes from a firm quote or a measured actual; grade B means it is derived by analogy from a similar part, project or cost-estimating relationship; grade C means it is an informed guess. The grade travels with the number into every total, dashboard and business case, so decision makers can see how much of a figure rests on evidence and how much on judgment. The discipline is simple: never add grades into a single total without showing the mix, and upgrade the C grades that drive decisions.
The method works because false precision is more dangerous than acknowledged uncertainty. A guessed $42,317 looks as solid as a quoted one on a spreadsheet. Grading makes the weak spots visible and focuses estimating effort where it changes a decision. It complements the AACE International estimate classification in 17R-97, which grades a whole estimate by the maturity of project definition, from Class 5 concept screening to Class 1 check estimates; the A/B/C grade works line by line inside such an estimate. Grades feed the Should-Cost Model, trigger Prototype Costing to upgrade critical lines, and are checked against reality in Variance Analysis.
What you need
- The cost estimate or business case broken down to line items
- Source of each number: quote reference, analog project, or estimator judgment
- List of decisions the estimate supports and their deadlines
- Budget and time available for upgrading estimates (quotes, teardowns, prototypes)
What you get
- Every line tagged A, B or C with its source
- Grade mix per module and for the total, by value
- List of decisions resting on C-grade numbers
- Upgrade plan for the critical C lines with owners and dates
- A trend of grade distribution across project phases
When to use it
When the business case adds a guessed cost to a quoted one.
How to do it, step by step
- Grade every cost estimate A (quoted), B (analogous), C (guessed).
- List the decisions that rely on C-grade numbers.
- Upgrade the critical C grades with quotes, teardowns or prototypes.
- Never add A and C grades into one total without showing the mix.
- Track the grade distribution per module over time.
Worked example: Grading a business case for a filtration skid product line
Illustrative scenario — figures are realistic but not from a real company.
An equipment builder prepared a business case for a new line of skid-mounted filtration units for industrial wastewater. The estimated unit cost was $186,000, and the approval committee wanted to know whether the target margin was realistic.
- The cost engineer graded 64 line items. Pumps, instruments and valves were quoted (A), the frame and piping were scaled from a previous skid (B), and the control panel software and the membrane housings were guesses (C).
- By value, the mix was 48% A, 31% B and 21% C. The C portion, about $39,000, sat mostly in two items.
- Because the margin decision hinged on the membrane housings, the team requested two supplier quotes and ran a quick teardown of a competitor housing.
- The software estimate was upgraded to B by analogy with the hours logged on two recent control projects.
Result. The housings came in $7,500 above the guess and the software $4,000 below. The revised unit cost of $189,500 left the margin slightly under target, and the committee approved with a design-to-cost action on the housings rather than discovering the gap after launch. The grade mix moved to 66% A, 30% B, 4% C.
Common pitfalls and how to avoid them
- Grading by estimator confidence rather than by source of the number.Define grades by evidence: A needs a document (quote, invoice, measured actual), B needs a named analog, everything else is C.
- Reporting only the grand total, which hides a heavy C share.Show the grade mix by value next to every total and on every dashboard line.
- Trying to upgrade every C line, which burns time on items that do not matter.Upgrade only the C lines that are large or that could flip a decision; leave small ones as C with a contingency.
- Letting old quotes keep their A grade after they expire.Give quotes a validity date and downgrade them to B once expired or when volumes change.
Frequently asked questions
What are the AACE estimate classes?
AACE International Recommended Practice 17R-97 defines five estimate classes based on the maturity of project definition, from Class 5 (concept screening, very wide accuracy range) to Class 1 (check estimate or bid, narrow range). Industry-specific practices, such as 18R-97 for process industries, give typical accuracy ranges per class. The classes grade a whole estimate; A/B/C grading applies inside it, line by line.
How accurate is an analogous cost estimate?
It depends on how close the analog is and how carefully it is adjusted for size, material, complexity and date. A near-identical previous part adjusted for inflation can be quite reliable; a loose analog from a different product family can be far off. Record the analog used and the adjustments made, then check the result against actuals through variance analysis to learn your real accuracy.
Should you add contingency to guessed costs?
Yes, but make it explicit. Carry C-grade lines with a visible contingency sized to the plausible range, rather than padding the number silently. That way the contingency can be released as the line is upgraded to a quote, and reviewers can see which part of the total is evidence and which is allowance.
Origin
Estimate classification — AACE International Recommended Practice 17R-97.
Used in these playbooks
Target setting workshop 1 h
One hour to set a defensible target cost: top drivers, scenario window, business case check — with confidence grades on every number.
- Top-Down Cost Drivers
- Cost Window Allocation
- Cost Business Case
- Estimate Confidence Grading
Related methods
- Should-Cost ModelBuild the bottom-up cost the product should have: materials at market, process times at standard, plus…
- Prototype CostingCost the prototype build honestly: it reveals the real process costs the estimate guessed.
- Cost Variance AnalysisExplain every gap between estimate and actual: which assumption was wrong, and what does that teach the model?
More in “Evaluate”
Grade the estimates and price the risks before deciding.