How to Evaluate Waste Management Software: A Buyer’s Framework

Waste management software is an unusually difficult category to evaluate, because most of the products in it demo well and differ underneath in ways a demo will not reveal. Every vendor will show you a dashboard with spend by site, a diversion rate, and a carbon number. The dashboards look similar. What differs is where each number came from, how much of it is measured versus inferred, who owns it, and who is expected to act on it — and none of that is visible from the screen.

This post is a framework for the questions that actually separate platforms. It is written by a vendor in the category, which you should factor in; we have tried to make it useful anyway by stating the questions in a form that a competitor could answer well, and by being explicit about the situations where a different kind of vendor — or no vendor — is the right answer. Those situations exist and pretending otherwise wastes everyone’s time in a sales cycle.

What should I actually be evaluating in waste management software?

Eight dimensions cover nearly everything that matters. They are not equally weighted for every buyer, and part of the exercise is deciding which three you care most about before you sit through a demo — otherwise you will be led through the vendor’s preferred three.

Where does the vendor’s data actually come from?

This is the first question and it subsumes several others. Ask the vendor to walk through one specific number on the dashboard — say, last month’s landfill tonnage for one site — and trace it backward to its origin. There are only three possible origins and they are not interchangeable.

The point is not that estimated data is illegitimate. Every large portfolio has some, and a platform that refuses to estimate anything will simply have gaps instead. The point is that a platform should tell you which is which, per number, without being asked. If it cannot, then any downstream figure — cost per ton, diversion rate, Scope 3 Category 5 — inherits an unknown amount of assumption and you have no way to size it.

Who owns the data, and can I get it out?

Waste platforms accumulate something genuinely valuable over a few years: a normalized, multi-year, multi-site record of what your operation generates and what it cost. Whether that asset belongs to you in a practical sense — not just contractually — determines your position at every renewal.

Three specifics to press on. First, export scope: can you pull the full record set at the row level, including the raw source documents the platform parsed, or only formatted reports? Formatted reports are not your data; they are the vendor’s interpretation of it. Second, export mechanism: is there a documented API, or is export a support ticket someone fulfills manually within a business week? Third, exit: what happens to the historical record if you leave, and is that written into the agreement rather than described verbally.

A useful test is to ask for a sample export during evaluation, on real data if a pilot allows it. Vendors whose export is genuinely first-class will hand it over; vendors for whom export is a retention lever will find reasons to defer.

What share of my invoices and contracts can it actually read?

Ingestion coverage is where stated capability and delivered capability diverge most, because the long tail of vendors is genuinely hard. A national hauler with an EDI feed is straightforward. The regional recycler that emails a scanned PDF with handwritten annotations, the grease collector that invoices quarterly with no site identifier, and the municipal service billed on a utility statement alongside water are not.

Ask three things. What ingestion channels exist — API, EDI, email parsing, PDF OCR? What happens to a document that fails to parse: does it silently drop, appear in an exception queue, or get handled by a human at the vendor? And what fraction of your specific vendor list, which you should provide, is already supported today rather than supportable in principle?

The normalization work behind ingestion matters as much as the parsing. Reconciling billing periods onto a common calendar, converting units, detecting duplicate invoices, and mapping vendor account numbers to your own site hierarchy is the unglamorous work that determines whether the portfolio is comparable at all. Dyrt does this in Spend Intelligence and writes about the mechanics in a separate post; the reason to raise it here is that it is easy to under-scope in an evaluation because no demo shows it.

How is the emissions number calculated?

This is the dimension where vendors most often present incomparable numbers as equivalent, and it is worth understanding well enough to ask a precise question.

The GHG Protocol Corporate Value Chain (Scope 3) Standard sets out, for Category 5 — waste generated in operations — a set of calculation methods that differ in how much they rely on assumption: a supplier-specific method using data from the waste treatment company, a waste-type-specific method using tonnage by waste type and treatment pathway, and an average-data method that leans on generic factors where activity data is missing. The methods are described in the standard itself. Two vendors can both claim GHG Protocol alignment while sitting at opposite ends of that range.

So ask which method, and then ask which factor set and which version. In the United States the common choices are EPA’s WARM, which compares management pathways on a material-specific life-cycle basis, and the EPA GHG Emission Factors Hub. Both are updated periodically, which means the version used is part of the answer: a restated prior year is a normal consequence of a factor update, and a vendor who cannot tell you which version produced last year’s figure cannot support a restatement.

The practical consequence of methodology choice is auditability. A figure built from weighed tonnage by material and destination can be traced to a chain of custody; a figure built from generic averages applied to spend cannot, and will not survive a serious assurance process. Dyrt’s Impact Engine takes the weighed-tonnage-plus-WARM route, described in the post on measured Category 5. That is a legitimate choice and not the only one — an average-data approach is explicitly permitted, cheaper, and adequate for some reporting obligations. What is not legitimate is presenting one as if it were the other.

Who does the work — my team or the vendor?

Analytics-only platforms share a characteristic failure mode that has nothing to do with the quality of their analysis. They produce an accurate, well-formatted list of opportunities, deliver it to a facilities organization that is already fully committed, and the list does not get worked. Twelve months later the identified savings are still identified.

Executing a single right-size means contacting the hauler, confirming the service change, verifying it appears on the next invoice, and catching it when it silently does not. Each instance is small; across a portfolio, in aggregate, it is a job. So the question to ask is not whether the platform surfaces findings — they all do — but who performs those steps, and what the vendor’s answer is when a change is confirmed by the hauler and then never appears on a bill.

Three service models exist and each is appropriate for someone: software-only, where your team executes; software plus advisory, where the vendor recommends and coaches; and fully managed, where the vendor holds the vendor relationships and executes on your approval. Dyrt’s tiers — Track, Capture, and Optimize — map to roughly that progression. The mistake buyers make is choosing a software-only product while implicitly assuming someone will have capacity to work the queue.

What does implementation actually require from my team?

Ask for the implementation burden in site hours, not vendor hours, and ask it separately for three phases.

If hardware is part of the proposal, the implementation questions change. Where does it physically go, what does the site provide, does it survive a receiving area with forklifts and moisture, and what happens during a network outage? A capture device that stops recording when connectivity drops will have its worst gaps at the sites with the worst infrastructure, which introduces a bias correlated with site type. Dyrt’s DWIT kiosk queues locally and reconciles on reconnect for that reason; the general point is to ask any hardware vendor what their offline behavior is, and to be skeptical of an answer that assumes reliable back-of-house WiFi.

When is a different kind of vendor the right answer?

Being honest about this is partly self-interested — a bad-fit customer is an expensive customer — but mostly it is that the alternatives are genuinely better in specific, identifiable situations.

There is also a compliance case that cuts the other way and is worth naming. Where a jurisdiction imposes specific operational obligations — California’s SB 1383 framework, for instance, requires commercial businesses to provide organics and recyclables containers alongside disposal containers and to monitor for contamination under 14 CCR 18984.9 — the requirement is operational rather than analytical. Software helps you evidence compliance; it does not create it. A platform that markets itself primarily as a compliance solution for a mandate that is fundamentally about containers and staff behavior is overselling.

What questions should I ask in the demo?

A short list that is hard to answer well without a real product behind it. Each one is answerable by a good vendor of any kind, including ones that are not us.

1. Pick a number on this dashboard. Trace it back to its source document or measurement for me.
2. Show me where the interface distinguishes a measured value from an estimated one.
3. Here is my vendor list. Which of these can you ingest today, and what is the fallback for the rest?
4. Which GHG Protocol Category 5 method do you use, which factor set, and which version?
5. When a right-size is approved, who calls the hauler, and who checks the next two invoices?
6. What does each site have to do in hours during the first ninety days, and every month after?
7. Export the full record set for a pilot site at row level, including source documents.
8. What causes this price to go up over three years?

Waste is a small line on most budgets and a disproportionately messy one, which is why it survives so long unexamined. Any of the vendor categories above will beat leaving it unexamined. The framework above is meant to get you to the right one rather than to us — and if the honest read of your situation is a broker, an ESG platform, or a spreadsheet, that is a legitimate conclusion from a good evaluation.