Most people evaluate a publisher in the wrong order. They open a metrics tool, look at a two-digit authority score, and only then glance at the site itself. We do it the other way around, and it saves an enormous amount of money. A score can be built. An archive cannot.
Here is the sequence we actually use when an Austrian publication is proposed for a campaign, and what each step is really asking.
Start with the archive
Open the publication and read backwards. Not the homepage, which is always the most presentable part of a site, but the category pages and the archive from two or three years ago.
You are looking for continuity. Did this site publish regularly before anyone was selling placements on it? Do the older pieces have the same tone and the same standard as the recent ones, or does the quality fall off a cliff at the point where the business model changed? A real publication has a past that looks like its present. A site built to sell links usually has a thin, recent archive with a sudden burst of commercial-sounding articles.
Bylines matter here too. Named authors who write more than one piece, who have a history elsewhere, who occasionally write something with no commercial angle at all. If every article is attributed to Redaktion and nothing else, that is not disqualifying on its own, but it removes one of the few signals you had.
Austria gives you one check that other markets do not. Sites operating here are required to carry an Impressum identifying who is behind them, and publications that see themselves as media usually go further and name an editorial contact and an ownership structure. Open it. A publication with a registered company, a named responsible person and a plausible address in Austria is a different proposition from one whose disclosure page is a generic template, an address in another country, or missing entirely. It takes thirty seconds and it settles more questions than any metric will.
Who is the page written for
Read three articles properly, in German. Not scanned, read. Ask yourself who the intended reader is. A piece written for an Austrian audience assumes things about that audience: local institutions, local terms, the Austrian version of a process. A piece written for nobody in particular hedges everything and names no institution at all.
Watch for translated content. Text that came out of English or out of a German original has a particular rhythm, and once you have noticed it you cannot stop noticing it. Sentences are slightly too long, idioms land slightly off, and the vocabulary is standard-German rather than Austrian. If the publication cannot be bothered to write for the market it claims to serve, its readers have probably already worked that out.
Pick any article and ask what the publication would lose if it deleted the piece. If the answer is nothing, because the page exists to hold a link rather than to inform anyone, you have learned what the site is for.
What the outbound links admit
Outbound links are the most honest part of a website. Editors link to sources. Placement shops link to customers.
So look at where a site points. A publication that cites statistics offices, courts, universities, other newsrooms and the occasional company because the company is the subject is behaving like a publication. A site where most external links are commercial, unrelated to each other, and pointing at categories the publication does not otherwise cover is telling you exactly how it pays for itself.
Pay attention to the verticals as well as the volume. If a general-interest Austrian site is linking into betting, adult content or short-term lending, your brand does not belong in that neighbourhood regardless of how the site looks on the surface. We treat that as an immediate no, and it is one of the tests we described on the homepage as non-negotiable.
Where the readers come from
A publication with an audience behaves differently from one without. You want evidence that people arrive from search, that they arrive from Austria, and that they arrive on more than one page.
Third-party traffic estimates are rough, and anyone who quotes them to three significant figures is overselling. Use them for shape rather than precision. Is the trend flat, rising or collapsing? Is the geography plausibly Austrian, or is most of the estimated audience somewhere the publication has no reason to be read? Does the site rank for anything a real reader would search, or only for its own name?
A publication that ranks for nothing has usually been penalised, deindexed in parts, or never earned anything in the first place. Any of those is a reason to walk away.
Check the indexation of the specific section your piece would appear in, not just the homepage. It is common enough to find a site where the news section is healthy and the guest-contribution category has quietly dropped out of the index, which tells you what Google made of the arrangement. If the pages that carry paid placements are the ones missing, the publication has already been judged for you.
What authority scores measure
Domain-level authority scores from the major tools are models of a link graph. They estimate how a domain is linked, nothing more. They do not know whether anybody reads the site, whether the content is any good, whether the audience is Austrian, or whether the links behind the score were bought in bulk last spring.
Because they are models, they can be gamed, and in a small market they frequently are. The pattern is recognisable once you look at the referring domains behind a suspiciously high score: a cluster of sites in unrelated languages, all linking sitewide, all acquired in a short window, none of them with an audience of their own.
Metrics still have a use. They are a fast way to sort a long list, and a sharp drop or spike is worth investigating. But they are the first filter, never the decision. We look at them after the archive, the audience and the outbound links, by which point they mostly confirm what reading the site already told us.
A high score on a site nobody reads is a number you are buying. A citation on a publication your customers recognise is an asset.
The order we run the checks in
- Is it Austrian in substance, writing in German for readers who live here, rather than hosted here or ending in .at.
- Does the archive have a past, with consistent publishing before any commercial model appeared.
- Would an editor defend the piece your link would sit inside, on its own merits.
- Is the outbound neighbourhood clean, with no high-risk verticals and no pattern of unrelated commercial links.
- Is there evidence of real readers, arriving from Austrian search on more than one page.
- Do the metrics agree with everything above. If they do not, the reading wins.
A prospect that fails any one of these is out rather than pending. If you want the version of this that ends with a shortlist on your desk instead of a checklist, that is what our services are, and the companion piece on what a natural Austrian link profile looks like covers what to do with the publishers once you have found them.