How we count
Zomodus sells an honest, neutral read of your customers' feedback. This page is that promise made checkable: exactly what a report is computed from, how each number is worked out, and the safeguards that stop anyone — the business or us — from bending the result. The headline number on every scan and report is the Z-ECM Score, defined below. Every report links here.
Last updated: Methodology v1.0 · 13 August 2026
On this page
1. What the report is computed from
Only your own customers' raw responses, run through one deterministic engine. No benchmarks, no invented data, no hand-editing.
A report is built entirely from the responses your own customers submit — nothing else. There are no industry benchmarks mixed in, no sample data, and no figures invented to fill a gap.
The same engine computes every report, the same way every time. Given the same responses it always produces the same numbers — so the result is a function of the feedback, not of who is looking or when.
The public demo report at /demo runs on this identical engine, so what you see there is exactly how your real report is produced.
2. The Z-ECM Score — the one number
Every Zomodus result leads with the Z-ECM Score: 0–100, in four plain bands, computed the same way for the review scan and your survey reports — so you track one metric from the very first scan onwards.
Z-ECM is short for the Z-Empirical Consistency Metric — Zomodus's single headline score. It runs from 0 to 100 and sits in one of four bands: Strong (88 and above), Solid (78–87), Mixed (68–77), or Slipping (below 68). The same score and the same bands appear on the instant review scan and on your ongoing survey reports, so the number never changes name or scale as you move from a first look to measuring properly.
On the review scan, the Z-ECM is computed from the public review ratings across your outlets — each outlet scored on the 0–100 scale, then combined into a group score weighted by how many reviews each outlet has, so a busy flagship counts for more than a quiet corner store.
On a survey report, the Z-ECM is the composite of what your own customers told you — overall satisfaction, the rated drivers, the recommend score, and the sentiment of open comments — expressed on that same 0–100 scale. The individual measures in the next section are the components that roll up into it.
3. How the numbers are worked out
The components behind the Z-ECM — plain averages and standard measures, no weighting tricks. Scores are averages; the recommend score is the usual NPS; priorities come from where low scores and negative comments line up.
Overall score and each rated item are the straightforward average of the 1–5 ratings customers gave, shown to one decimal place.
The recommend score (NPS) is the standard net promoter measure: the percentage of promoters (9–10) minus the percentage of detractors (0–6), on a −100 to +100 scale.
“What to fix first” is ranked by combining where scores are lowest with where written comments are most negative — so the report points at the real weak spot, not the prettiest chart. Confidence is flagged honestly: fewer responses means a more cautious read.
4. The business cannot game its own score
Responses are write-once. A business can never add, edit, or remove a customer's response, and repeat submissions are blocked before they can pad the result.
Once a customer submits, their response is recorded permanently and is never editable or deletable by the business. There is simply no screen anywhere that lets a business change or drop its own feedback.
Repeat and duplicate submissions are stopped before they are recorded: each device can answer a given survey once per period (for a monthly survey, once per calendar month), so no one can stuff the results by submitting again and again.
Because the business owns neither the counting nor the record, the score it sees is the score its customers gave — which is the whole point of a neutral, third-party read.
5. A tamper-evident record
Every response carries a cryptographic fingerprint chained to the one before it, so altering, reordering, or deleting any past response would be detectable.
As each response is recorded, Zomodus computes a SHA-256 fingerprint (a hash) of its content, the time it was recorded, and the fingerprint of the previous response for that survey. Chaining each response to the last one turns the whole set into a sealed sequence.
If any past response were changed or removed after the fact, the chain would no longer line up — the tampering would show. This is the same idea that secures other trusted digital records.
When you submit a response you are shown a short reference code drawn from that fingerprint. It lets your specific submission be pointed to later without revealing what you said.
6. Your privacy is not the price of proof
We prove integrity with counts, timestamps and fingerprints — never identities. Responses stay anonymous, in line with our privacy policy and the PDPA.
None of the safeguards above rely on knowing who a respondent is. Integrity is proven with counts, timestamps and one-way fingerprints — not names, contacts, or tracking.
Feedback stays anonymous to the business. Proving a number wasn't gamed and exposing who gave it are two different things, and Zomodus only ever does the first.
7. Independent verification (coming next)
We are extending this so a third party — an auditor, or a buyer's diligence team — can verify a report's record independently, without taking our word for it.
The safeguards above already make the record tamper-evident. The next step, in progress, is a signed export of the response record that anyone can check independently, plus a way to confirm a published report hasn't been altered since.
This methodology is versioned. When how we count changes in any material way, the version above changes with it, and every report states the version it was produced under.