Publisher discovery attracts inflated claims: invented traffic numbers, scraped contact lists, precision figures nobody measured. We publish what we do — and what we refuse to do — so you can hold us to it.
These are not aspirations — they are properties of the pipeline. Each one is visible in the file you receive.
Every extracted signal carries its own confidence score. You can filter to high-certainty rows, or dig into the uncertain tail — but you always know which is which.
Full runs start only after precision is measured on a sample your team has reviewed. The number we quote is the number you helped verify — not a marketing figure.
Traffic is described through signals anyone can re-check: presence in the Chrome UX Report and rank groupings. Proxies are honest; estimates dressed as measurements are not.
Every signal and custom field is defined in writing before the run. If a definition changes during calibration, the change is documented — nothing shifts silently.
Deliveries are deduplicated and canonicalized: mirrors, redirects, and variants collapse to a single canonical domain. Counts you pay for are counts of real, distinct sites.
We are a Munich-based EU company and build deliveries around publicly published site characteristics. Data minimization is the default, not an afterthought.
Some things are standard practice in this market. We decline them, and we say so before you ask.
We do not harvest and resell personal emails or phone numbers. The deliverable is domains and site-level signals; your outreach uses channels the publishers themselves publish.
You will not find "monthly visitors: 48,200" in our files, because nobody outside a site's own analytics knows that number. We report reproducible proxies or nothing.
We describe reference engagements by category and numbers only — never by name. The same discretion protects you when you become the reference.
Public sample material is paraphrased and anonymized: real structure, real field logic, altered identities. What we show never exposes a client's actual deliverable.
Precision figures we cite come from reviewed samples on real engagements. If a vertical hasn't been calibrated yet, we say "we'll measure it in your pilot" — not a made-up percentage.
No fabricated testimonials, no logo walls of companies who never heard of us. Our sales pitch is the free pilot — data on your vertical, judged by your team.
Every column in a delivery answers three questions: where does it come from, how sure are we, and how was it verified?
| Claim type | Source | Verification |
|---|---|---|
| Industry relevance | Classification database + LLM read of the live site | Reviewed-sample calibration before the full run |
| Publisher signals | LLM extraction from published site content | Confidence score per signal, spot-checked in calibration |
| ICP fit | Computed from required / preferred / disqualifying signals | Logic agreed with you in writing before the run |
| Traffic proxy | CrUX presence, rank group | Reproducible by anyone from public sources |
| Domain identity | Dedup and canonicalization pass | One row per canonical publisher, variants mapped |
The free pilot delivers 500 qualified publisher domains with every confidence score visible — judge the quality yourself before spending anything.
Request a Free Pilot