What Should Waste Cost Per Location? Why Published Averages Cannot Answer That
It is one of the first questions a CFO asks when waste finally gets attention, and it is the right question to ask. If we run 340 locations, what should waste cost at each one? Give me the number and I will tell you whether we have a problem.
The uncomfortable answer is that no external number will settle it. Not an industry average, not a peer figure from a conference slide, not a per-square-foot rule of thumb. Waste pricing is set locally, by market structure and regulation and haul distance and what your site actually throws away, and the variance across those factors is wide enough to swallow any average you could compute. A national mean for cost per location is a real number in the sense that arithmetic was performed. It is not a benchmark, because it cannot tell you whether a specific site is expensive.
The benchmark that works is one you build from your own portfolio. This post is about why the published version fails and how to construct the internal version so it survives contact with a hauler.
Why is there no reliable average waste cost per location?
Start with the fact that commercial waste is not one market. It is thousands of local markets operating under incompatible rules, and in some of them the price is capped by a regulator while in others it is whatever two parties agree to.
In New York City, the Business Integrity Commission sets a ceiling on what private carters may charge for putrescible and recyclable waste removal. As of the effective date published on the BIC maximum rates page, that ceiling is $26.87 per cubic yard of loose refuse by volume, or $17.64 per 100 pounds by weight, with the customer and carter free to negotiate which basis applies and free to agree to anything below the cap. Medical waste, yellow grease, e-waste, construction and demolition debris, and paper collected for shredding are explicitly outside the cap.
Now cross the country. Portland, Oregon permits commercial haulers but does not price them. The city states the position plainly on its permitted commercial garbage and recycling companies page: businesses can negotiate service level, frequency and cost, and rates are not set by the City.
Two sites in the same brand, with the same footprint and the same menu, sit inside those two regimes. One has a published legal ceiling denominated in cubic yards or hundredweight. The other has whatever the account manager quoted. Averaging them produces a figure that describes neither. Add a third site in a franchised municipal system where a percentage of gross revenue is levied on the hauler and passed through — the Derby, Kansas solid waste RFP asks bidders to address a franchise administration fee of five percent of gross revenue — and the average is now mixing in a tax.
What is actually different between two sites that look the same?
Regulatory regime is only the first divergence. Once you start pulling invoices apart, cost per location decomposes into a set of drivers that vary independently of each other, which is precisely why a single aggregate number cannot be reverse-engineered into a diagnosis.
Service configuration is the driver finance teams underestimate. A site paying for a six-yard container emptied five times a week and a site paying for a thirty-yard compactor pulled when full are buying fundamentally different products, and the second one may be cheaper per ton while looking more expensive per month. Stream mix compounds it: a location with a food and beverage operation generates organics and higher-density waste, and in California that is not merely an operational fact but a compliance obligation, since the state organics requirements described by CalRecycle under SB 1383 obligate jurisdictions to provide organic waste collection to residents and businesses. A mandated third stream is a third set of containers, pulls, and line items that a comparison site in another state simply does not have.
Drivers that move cost per location independently of each other
| Driver | How it moves cost | Why it breaks external averages |
|---|---|---|
| Regulatory regime | Rate cap, franchise fee, or open negotiation | A capped market and a free market are not comparable observations |
| Disposal market | Tip fees and haul distance to the nearest facility | Set by local landfill and transfer capacity, not by your vendor |
| Service configuration | Container size multiplied by pickup frequency | Two sites can pay the same monthly total at very different unit rates |
| Stream mix | Number of separated streams and their densities | Mandated organics adds a stream some comparison sites lack entirely |
| Property type and access | Compactor versus open top, dock constraints, gated access | Physical access dictates equipment, and equipment dictates price |
| Contract vintage | Years of compounded escalation since signature | An older agreement at the same headline rate now bills higher |
Contract vintage is the driver nobody normalizes for
Two sites can hold the same contract, at the same original rate, from the same vendor, and bill differently today purely because one was signed four renewal cycles earlier. Unless your comparison controls for how long a rate has been escalating, you will read a paperwork artifact as an operational problem.
What would a national average have to hold constant to be useful?
Work the question backwards. For an external per-location figure to tell you anything about your site, the sample behind it would need to match your site on regulatory regime, disposal market, container and frequency configuration, stream count, property type, and contract age. Nobody publishes a figure conditioned on six variables, because conditioning on six variables leaves almost no sample.
What is published nationally is material generation and management data at the country level. The EPA national overview of facts and figures on materials, waste, and recycling is a genuinely authoritative source, and it is authoritative about exactly what it measures: aggregate national tonnage and management pathways. It is not a per-site cost benchmark and does not claim to be. Most of the per-location dollar figures in circulation have no comparable provenance at all.
Where external data does earn its place is in measuring change rather than level. The Bureau of Labor Statistics publishes a producer price index for the industry — series PCU562111562111, titled PPI industry data for solid waste collection, not seasonally adjusted. That series will not tell you what your Phoenix site should pay. It will tell you whether the increase your vendor is proposing is in line with what happened to industry prices, which is a far more answerable question and a far more useful one in a negotiation.
How do I build an internal waste cost benchmark?
The defensible benchmark is your own portfolio, normalized well enough that sites can be held next to each other. That is a data problem before it is an analysis problem, and it has a specific sequence.
Every hauler, every site, every month, for at least four quarters. Partial coverage produces a benchmark that quietly excludes your worst sites, because the sites with messy billing are usually the ones with problems.
A hauler account number is not your property ID, and the mapping drifts as sites are acquired, renamed, or re-regioned. Until the join is reliable, no site-to-site comparison is trustworthy.
Haulers close months on different days and bill in tons, cubic yards, gallons, and per-pull. Convert to a common basis and a common calendar before computing anything per location.
Separate base service from surcharges, taxes, franchise fees, and event-driven charges. The effective all-in per-unit rate is the comparable figure; the monthly total is not.
Group by market, property type, stream configuration, and service configuration. Compare within cohorts. A cohort of four similar sites is more informative than a portfolio-wide mean.
An expensive site is a hypothesis, not a finding. The contract says whether the rate is permitted, whether the surcharge follows the agreed formula, and whether the escalation is the one you signed.
Steps one through four are the work that Spend Intelligence exists to do. It ingests invoices monthly through whichever path a vendor supports — APIs, EDI, email parsers pointed at an accounts-payable inbox, and PDF OCR for the scanned bills that make up more of the corpus than anyone expects — then reconciles billing periods across haulers, converts units to a common basis, deduplicates submissions that arrived through two channels, and maps vendor facility identifiers onto your site hierarchy. The mechanics are covered in detail in how Spend Intelligence reads a hauler bill.
The unit that makes sites comparable
Cost per location is the number leadership asks for, but it is not the number that diagnoses anything. Effective all-in cost per unit of material handled, computed within a cohort of genuinely similar sites, is what identifies which sites are actually mispriced. Report the first; act on the second.
What does the benchmark let me do that an average does not?
An external average produces an argument. An internal cohort comparison produces an action, because it names the site, the rate, the comparable, and the vendor who is charging both.
The findings that come out of a normalized portfolio fall into a small number of recurring shapes, and the correct response differs for each. Treating them all as "negotiate harder" is how programs stall.
Four finding types, four different responses
Site is priced above its own cohort
The effective all-in unit rate at this location is materially above what comparable sites in the same market pay the same or a competing vendor.
Recommendation: Take the internal comparison into the renewal. A vendor can dispute an industry average; it is harder to dispute your own invoices from its own trucks.
Configuration no longer matches demand
Container size and pickup frequency were set for a demand profile the site has since lost. Consistently light hauls are the tell, and they read as high cost per unit.
Recommendation: Reduce frequency or downsize before renegotiating rate. Buying less service beats buying the same service cheaper.
Cost gap traces to surcharges, not base rate
Base service is in line with the cohort but the all-in figure is not, because fees, taxes, and pass-throughs are carrying the difference.
Recommendation: Decompose the surcharge stack and require substantiation for each component against the contract and any applicable published tariff.
Billed events have no service record
Extra pickups, overweight charges, or contamination fees appear for a site with no corresponding record that the event occurred.
Recommendation: This is a credit conversation, not a negotiation. Package the evidence and file it while the charge is one cycle old.
Where the evidence needs to be stronger than an interpretation, measurement helps. A DWIT kiosk pairs a three-tap workflow with an industrial scale rated above 5,000 pounds, which produces a weighed and timestamped record you can hold against a billed haul. A cost-per-location figure derived partly from measured weight is a different class of number than one derived entirely from what a vendor asserted.
How often should the benchmark be rebuilt?
Continuously, because every input drifts. Disposal markets reprice, contracts escalate on their own schedules, sites open and close, and service configurations get changed at the property level by managers solving an operational problem without reference to cost. A benchmark computed once and referenced for three years is a historical document.
For portfolios past roughly fifty locations, keeping it current is a staffing question as much as a software question. Portfolio Intelligence is the version where Dyrt runs it: national vendor management, continuous invoice audit and dispute resolution, right-sizing and switch execution on your approval, ground-truth verification from loading docks and food and beverage teams, monthly reporting, and quarterly business reviews. That operating model is described in the Portfolio Intelligence post.
The honest version of the original question is therefore not "what should waste cost per location" but "which of my locations cost more than my own comparable locations, and why." That question has an answer, the answer is defensible in front of an auditor and a vendor, and the data required to produce it is already sitting in your accounts-payable inbox.
Dyrt Team
Dyrt Editorial
The Dyrt team builds waste intelligence software for sustainability managers, CFOs, and facility operators. We help organizations reduce waste costs, hit diversion targets, and simplify Scope 3 reporting.
Ready to Transform Your Waste Operations?
Dyrt helps sustainability teams, facility operators, and CFOs cut waste costs, hit diversion targets, and simplify Scope 3 reporting — all in one platform.
Request a Demo
