The short version
- The line is published. Google’s spam policies describe what is prohibited, and most borderline cases are addressed there directly.
- The practical test: would this make sense if search engines did not exist? Legitimate tactics usually would.
- The risk is asymmetric — a tactic that fails costs you the effort; a tactic that breaches policy can cost rankings you already have.
- The "grey hat" area is smaller than practitioners claim. Most of what gets called grey is clearly described in the policies.
This framing has been around since the early days of SEO and it has become vaguer over time, mainly because nobody selling a borderline tactic wants to call it black hat. The useful correction is that the policies are published, specific, and cover most of what people describe as ambiguous.
The clear cases
Very little on that list is genuinely ambiguous. The ambiguity people report usually comes from practitioners describing a prohibited tactic in language that makes it sound like a legitimate one — "relationship-based link acquisition" for paid links, "content at scale" for mass-produced thin pages.
The tests for a borderline case
- Would you do this if search engines did not exist? A useful article, yes. A paid link on an unread site, no.
- Would you explain it to the site you are working with? Practices that require concealment from your own partners are answering the question.
- Does it show users something different from crawlers? That is cloaking, and it is unambiguous.
- Is the value real or manufactured? A link from a site with real readers is earned. A link from a site that exists to sell links is bought.
- Would you still do it if the link were nofollowed? If no, you are buying a ranking signal — see guest posting vs digital PR.
Why the risk is worse than it looks
The asymmetry is what makes this a practical question rather than an ethical one. A legitimate tactic that does not work costs you the effort you spent. A prohibited tactic that gets caught can cost rankings you already had — and recovery is slow, uncertain and expensive.
That asymmetry gets worse the more successful you are. A site with no rankings has nothing to lose; a site earning real revenue from organic search is risking a functioning channel for a marginal gain. The calculation that made sense for an affiliate site in 2012 does not transfer to a business with customers.
Why the labels have stopped being useful
The hat metaphor implies a spectrum of intent, which is why it keeps getting stretched. A more useful framing is simply whether a tactic would survive being described accurately to everyone involved — the publication, the client, the reader.
That test is harder to argue around than "is this black hat", because it does not depend on where you draw an imaginary line. "We write useful articles for publications our customers read" survives description. "We pay for placements on sites selected by authority score and negotiate the anchor text" does not, and no amount of terminology makes it sound better to the person you would have to say it to. The published spam policies describe the second case directly.
The strongest argument against black hat is not ethics
It is durability. Tactics that exploit a specific weakness stop working when the weakness is fixed, and you rebuild from scratch. Work that makes your site genuinely more useful compounds and survives algorithm changes. One approach is a treadmill; the other is an asset.