Dental clinic, mid-size town
Score 31 → 68. Root cause: zero measurement on a four-year ad budget.
31
68
audit score
Read the case
VERIFIABLE RESULTS
Before and after, measured with the same method.
Every case keeps the same scoring framework and criteria. That is the difference between a marketing story and a change you can actually compare.
RESULTS
What changed, measured.
Every result below has a before score and an after score, computed with the same method.
Independent car service
Score 24 → 59. No form, no visible phone number on mobile.
24
59
audit score
Read the case
Niche online shop
Score 42 → 71. Pixel installed wrong, remarketing on empty data.
42
71
audit score
Read the case
Law firm
Score 38 → 64. Good content, completely invisible to engines.
38
64
audit score
Read the case
METHODOLOGY
Reproducible scores. AI with clear limits.
We separate measurement from interpretation. The scanner calculates, AI explains, and validation marks what lacks enough support.
1. Collection
Data comes from the form, the site, public research and available comparisons.
2. Measurement
The scanner calculates 0–100 scores from deterministic signals on the same scale.
3. Context
Research and competitors add external context when sources are available.
4. Interpretation
AI explains impact and orders priorities without permission to change the scores.
5. Truth check
The report explicitly declares assumptions, missing live research or claims that cannot be supported strongly enough.
What actually gets measured
Every number in the report has evidence behind it. If something could not be verified, the report says so instead of covering it.
15
pages scanned
Not just the homepage. We open the important pages found in the menu.
6
deterministic scores
The same site gives the same score. You can compare before and after.
46
signals per page
Structure, forms, calls to action, structured data, tracking, local signals.
0
unverified claims
What cannot be confirmed is marked as an assumption, not presented as fact.
WHAT CLIENTS SAY
No superlatives.
The first time anyone showed me concretely what was broken, instead of what I should buy. The report had screenshots from my own site.
They told us outright that two claims in the report were assumptions, because they had no external data. I had never seen an agency do that.
We ran the audit before and after. Same criteria, same score computed the same way. I could finally compare something.
TRANSPARENCY AND CONTROL
We say what we know and what we do not.
A useful report must be auditable. We keep measured signal, external context and interpretation separate.
No invented numbers
Displayed scores come from the scanner, not from an AI estimate.
Visible caveats
If external research did not run or a claim has weak support, the user can see that.
Abuse protection retained
Turnstile, nonce and the existing engine rate limit stay in the flow; the builder creates no alternate bypass.
Server-authoritative pricing
The browser sends only the package ID. The amount is read again from the server catalogue when the order is created.
Controlled data use
Analysed pages are public; protected areas are not accessed and the builder does not modify the audited website during scanning.
FIRST STEP
See your score before paying anything.
The analysis is complete and free. No card, no account.
Start the analysis