The right method is bottom-up estimating, built from a work breakdown structure that separates resin sourcing, bottle forming, tooling, and testing, then priced with a stated accuracy range of −10% to +15% at the Class 3 gate. Class 3 is the Association for the Advancement of Cost Engineering (AACE International) designation, and it suits this project because you know the bottle format and the recycled feedstock route, but not the final machine or supplier contracts. Skip top-down ratios here; food-contact compliance and recycled resin pricing will break a square-foot or per-unit shortcut.
Why Top-Down Ratios Fail on Recycled-Resin Bottle Projects
A per-bottle ratio pulled from virgin PET packaging lines will understate your costs in three places. Recycled feedstock carries a premium and a quality risk that virgin resin does not. Your line may need drying and decontamination equipment that a standard blow-molding cell doesn't include. And food-contact approval testing is a discrete cost block that no per-unit ratio captures.
The premium is real and it is recorded. The National Association of PET Container Resources has tracked food-grade recycled PET (rPET) trading well above virgin PET resin in most quarters since 2021, with the gap widening when collection volumes fall. A ratio built on virgin resin pricing therefore misses the single largest line in your material budget.
Consider a concrete mismatch. The U.S. Plastics Pact, coordinated by the Recycling Partnership, set a target of 30% recycled content in plastic packaging by 2025 across its signatory brands. Any bottle project built for those customers needs rPET volume that must be contracted in advance, and contract terms include minimum tonnage and quality specs that fixed ratios ignore.
Top-down estimating does have a place: the AACE Class 5 gate, where a ±30% to −50% range is acceptable for a screening decision. Use it only to decide whether the project deserves a budget at all. Once someone is writing checks for molds, switch methods.
The Work Breakdown Structure You Need Before Any Number Goes In
Build the WBS before pricing anything, because a missing branch is a missing cost that no contingency can rescue. For a recycled-material bottle, five branches cover the project: feedstock, conversion, tooling, compliance, and launch logistics. Each branch gets its own estimator and its own accuracy note.
Feedstock covers resin grade selection, bale or flake sourcing, contracts, incoming inspection, and drying or decontamination. Conversion covers the forming process, whether that is extrusion blow molding, injection stretch blow molding, or a thermoforming route for rPET sheet. Keep those separate; they carry different machine capital.
Tooling is its own branch because molds and blow pins are long-lead and priced per cavity. Compliance covers food-contact testing, migration testing, and any third-party certification the customer demands. Launch logistics covers packaging, freight, and the scrap rate during line qualification.
A 20-work-package WBS is workable for a single bottle format. Bottom-up estimating on fewer than about 15 packages tends to hide costs inside a generic "equipment" line, which is exactly where budget overruns breed.
How Each Cost Element Is Actually Priced
Price material from a contract quote, not a spot index, because rPET contracts set minimum tonnage and quality specs. Ask for a landed price per kilogram including freight and any blending. If you can only get an index, apply the AACE guidance that material taken from public indices carries wider accuracy than a supplier quote.
Price conversion from a machine-hour rate you build yourself. Take the equipment cost, add installation and utilities, divide by realistic annual operating hours, and load labor and overhead. Three-shift operation at roughly 6,000 hours per year is a common planning figure for a dedicated cell; confirm your own plant's calendar before you use it.
Price tooling from written mold-maker quotes with cavity count and cycle time stated. A single-cavity prototype mold and a 16-cavity production mold differ by an order of magnitude, and the quote should name which you are buying. Price compliance testing from a named laboratory's scope of work.
Price scrap and yield loss explicitly. Recycled resin runs at different melt properties than virgin, so your first production runs will scrap more bottles. Put a learning curve in the estimate rather than hiding it in contingency.
| Cost element | Pricing source | Typical accuracy driver |
|---|---|---|
| rPET feedstock | Supplier contract quote | Grade, tonnage commitment, freight |
| Machine time | Built-up hourly rate | Utilization hours, labor load |
| Molds and tooling | Mold-maker written quote | Cavity count, cycle time |
| Compliance testing | Named lab scope of work | Test standard, sample count |
| Yield loss | Historical or pilot data | Resin variability, line maturity |
Setting the Accuracy Range: Class 3 and What −10% to +15% Buys You
State your accuracy range on the estimate cover sheet or the number will be treated as a promise. AACE International's cost estimate classification system defines Class 3 as a range of −10% to +15%, and it pairs that with roughly 30% to 75% of project definition complete. That is the normal gate for a capital appropriation request.
Class 3 assumes most of the engineering is identified but not detailed. For a bottle project that means you have a chosen forming route, a named resin grade, and a preliminary plant layout, but not final drawings or signed equipment contracts. If any of those three is still open, you are at Class 4 or 5 and should not present Class 3 accuracy.
Understating the range is the most common estimating failure in packaging projects. Presenting a single-point number invites the finance team to hold you to it, and the +15% upper bound exists because tooling and compliance testing routinely land above plan.
An accuracy range is not contingency. Contingency covers identified risk; the range describes how uncertain the estimate itself is. Keep them in separate lines so reviewers can see both.
When to Use Parametric or Analogous Estimating Instead
Parametric estimating fits when you have a genuine cost driver with historical data behind it, such as cost per cavity per mold set, or cost per million bottles produced. A parametric model is only defensible if your own plant has run similar lines and you have actuals to regress against. For a first recycled-resin line, you probably don't.
Analogous estimating fits early screening and nothing more. Take the cost of a comparable bottle project, adjust for capacity and material, and label the result as Class 4 or 5. It gets you into a budget conversation; it does not get you an appropriation.
Hybrid approaches are common and legitimate. Price tooling and compliance bottom-up because they are discrete and quote-driven, then use a parametric rate for conversion because machine time scales predictably with cavity count and cycle time. Blend the two and state which portions carry which accuracy.
Recycled-material projects break parametric models more often than virgin-resin projects do, because feedstock variability changes cycle time and scrap rate. A model built on steady-state virgin resin will drift low on scrap and high on throughput assumptions.
Procurement and Feedstock Risk That Changes the Number
Recycled resin availability is the swing factor most estimates miss. Contract for feedstock before you finalize the capital estimate, or price a feedstock risk allowance into the number. Collection rates, bale quality, and competing demand from beverage brands all move the contract price.
Quality specification drives cost as much as price does. Food-grade rPET requires decontamination and certification to a recognized standard, and the equipment to meet it is a capital cost. If your customer's spec is tighter than standard food-contact, add the incremental testing and processing.
Volume commitments tie up working capital. A supplier may require minimum annual tonnage to hold a price, and that tonnage sits in inventory until your line consumes it. Working capital belongs in the project's funding request even though it isn't a capital equipment cost.
Single-source risk is a cost, not just a risk register entry. If only one supplier can meet your grade, budget for a second qualification run and possibly a smaller-volume secondary contract at a worse price. That contingency is real money.
A Worked Two-Line Estimate Showing the Range in Dollars
Use a transparent arithmetic example rather than a borrowed benchmark. Assume two blow-molding lines, each running 6,000 hours per year, producing 2,000 bottles per hour per line. That is 24 million bottles annually at full utilization. Price the estimate in three blocks: equipment, tooling and compliance, and first-year operating cost.
Take equipment at a round planning figure of 2.4 million dollars for two lines including installation and utilities. Take tooling at 320,000 dollars for two 16-cavity production molds plus a prototype mold. Take compliance testing at 90,000 dollars for a full food-contact test program with a named lab.
For operating cost, assume feedstock at a contract price and yield loss of 4% in year one, dropping to 2.5% in year two. At 24 million bottles and an assumed bottle weight, the material cost dominates. If a 20-gram bottle consumes 0.020 kilograms of resin, 24 million bottles need 480 tonnes of resin before scrap, and 500 tonnes at 4% scrap.
| Block | Planning figure | Accuracy band applied |
|---|---|---|
| Equipment and installation | 2,400,000 USD | −10% to +15% |
| Tooling and prototype mold | 320,000 USD | −10% to +15% |
| Compliance testing | 90,000 USD | −10% to +15% |
| Year-one feedstock at 500 tonnes | Set by contract price per tonne | Contract, not estimate |
Apply the band to the capital blocks: the low end of 2,810,000 dollars is about 2,529,000 dollars and the high end is about 3,232,000 dollars. Feedstock cost sits outside the band because it is a contract, not an estimate. Confirm your own bottle weight and contract price before you publish a number; this example is arithmetic, not a market quote.
Review Gates, Contingency, and Who Signs Off
Contingency should be sized from an identified risk list, not a flat percentage. List the risks: mold rework, failed migration test, feedstock price movement, line qualification overrun. Assign a probability and a dollar impact to each, then sum the expected values.
Review the estimate at three points: after the WBS is complete, after supplier quotes are in, and after the pilot run. Each review should either tighten the accuracy range or explain why it hasn't tightened. An estimate that never narrows is an estimate nobody is managing.
Sign-off belongs to whoever owns the capital budget, with the estimator presenting the range and the risk list, not a single number. Ask the approver to accept the upper bound, not the midpoint. If they can't accept the upper bound, the scope needs cutting before approval.
Keep a cost history file from day one of the project. Every actual against estimate becomes a data point for the next bottle format, and that file is what eventually makes parametric estimating defensible in your plant.
Common Estimating Mistakes on Recycled-Content Packaging
The first mistake is pricing recycled resin at a virgin index because it's easier to find. Use a contract quote or state clearly that the material line carries wider accuracy than the rest of the estimate.
The second is burying compliance testing inside a general "engineering" line. Food-contact testing has a discrete scope and a discrete price; pull it out where reviewers can see it.
The third is omitting yield loss in the first year. Recycled resin behaves differently on the line, and the difference shows up as scrap. Put a learning curve in the model.
The fourth is presenting a midpoint as the budget. That invites a funding shortfall the moment tooling or testing comes in high. Present the range and make the approver choose a point within it deliberately.
The fifth is skipping the pilot run. A pilot produces actual cycle times and scrap rates, which converts an estimate into a forecast. Budget for the pilot as a project cost, not an afterthought.
FAQ
What estimating class should a bottle project be at when it goes to the capital committee?
Class 3, with a −10% to +15% range, is the normal gate for capital appropriation under AACE International's classification system. That class assumes roughly 30% to 75% of project definition is complete. If your forming route, resin grade, or plant layout is still undecided, you are at Class 4 or 5 and the range should be wider.
How much does food-grade recycled PET cost compared with virgin PET?
Food-grade rPET has traded above virgin PET resin in most quarters since 2021 according to National Association of PET Container Resources market reporting, with the gap widening when collection volumes tighten. The exact spread moves weekly, so confirm the current spread with your supplier before you commit a material line. Treat the premium as a range, not a fixed number.
Can I just use a cost-per-bottle ratio from an existing packaging line?
Only at a screening gate, and only if you label the result Class 4 or 5. A ratio built on virgin PET will miss the recycled feedstock premium, decontamination equipment, and food-contact testing. For appropriation, price those three items bottom-up and use the ratio for conversion only.
What happens if the migration test fails after the mold is cut?
You pay for mold rework or a resin reformulation, and the timeline slips. That is why the compliance block gets its own contingency line and why the pilot run should include a preliminary migration screen before production tooling is ordered. A failed test on a production mold is the most expensive compliance outcome in this kind of project.
Is bottom-up estimating always better than parametric?
No. Bottom-up costs more to build and takes longer, so it only pays off once the project is real. Parametric is faster and cheaper when you have historical actuals from your own plant to regress against. For a first recycled-resin line with no internal history, bottom-up is the defensible choice.
How many work packages should the WBS have for a single bottle format?
Around 20 work packages is workable for one format. Below about 15, costs hide inside generic equipment lines and overruns become invisible until too late. Each package should have one owner and one pricing source attached.
Does working capital belong in the project estimate?
Yes, when a feedstock contract requires minimum annual tonnage. That inventory sits until the line consumes it, so it belongs in the funding request even though it isn't capital equipment. Keep it as a separate line so the capital committee sees the full cash requirement.
What single number should I give the approver?
Give the upper bound of the range and the risk list behind it. Ask them to accept the upper bound, not the midpoint. If the upper bound is unacceptable, cut scope before approval rather than squeezing the estimate to fit.